作者Rupesh Raj Karn, Johann Knechtel, Ozgur Sinanoglu
As graph neural networks (GNNs) become standard tools for critical tasks in circuit design and analysis, their security and privacy risks require careful attention. Here, we present the first comprehensive evaluation of gradient leakage attacks (GLAs) on GNNs in circuit-design and hardware-security tasks, a practical threat that has been largely overlooked. We assess state-of-the-art (SOTA) GNNs, including GraphSAGE, GCN, GIN, and GAT, trained on standard netlist benchmarks (ISCAS'85, EPFL, and TrustHub), for their fundamental vulnerability to GLAs. We find that GLAs can expose sensitive information, such as gate types and distinctive properties of hardware Trojans, which may assist adversaries in analyzing logic locking schemes or evading Trojan detection mechanisms. Our analysis shows that these risks are influenced by architectural features, with attention mechanisms (GAT) exacerbating leakage, while injective aggregation (GIN) provides comparatively stronger resilience. We further evaluate several SOTA defense techniques, including differential privacy, gradient clipping, secure aggregation, model compression with quantization, and adversarial training. We find that these techniques improve resilience only in specific settings and can also compromise model performance. Overall, our work provides key insights toward privacy-preserving GNNs and highlights the need for more robust and efficient defenses. We release our full methodology and artifacts.
展开完整摘要收起摘要↓
As graph neural networks (GNNs) become standard tools for critical tasks in circuit design and analysis, their security and privacy risks require careful attention. Here, we present the first comprehensive evaluation of gradient leakage attacks (GLAs) on GNNs in circuit-design and hardware-security tasks, a practical threat that has been largely overlooked. We assess state-of-the-art (SOTA) GNNs, including GraphSAGE, GCN, GIN, and GAT, trained on standard netlist benchmarks (ISCAS'85, EPFL, and TrustHub), for their fundamental vulnerability to GLAs. We find that GLAs can expose sensitive information, such as gate types and distinctive properties of hardware Trojans, which may assist adversaries in analyzing logic locking schemes or evading Trojan detection mechanisms. Our analysis shows that these risks are influenced by architectural features, with attention mechanisms (GAT) exacerbating leakage, while injective aggregation (GIN) provides comparatively stronger resilience. We further evaluate several SOTA defense techniques, including differential privacy, gradient clipping, secure aggregation, model compression with quantization, and adversarial training. We find that these techniques improve resilience only in specific settings and can also compromise model performance. Overall, our work provides key insights toward privacy-preserving GNNs and highlights the need for more robust and efficient defenses. We release our full methodology and artifacts.
Large-scale neural-network training repeatedly aggregates gradients across devices, making communication a central cost in distributed learning. Low-bit gradient aggregation can reduce this cost, but applying it as a static replacement for full-precision communication can destabilize training because safe precision depends on training phase, model structure, runtime bucketization, and the communication substrate. This paper presents NEURON-Fabric, a profile-guided runtime system for controlled low-bit gradient communication. NEURON-Fabric uses calibrated operating profiles, model-aware runtime bindings, online training-health monitoring, and reducer-capacity checks to decide when low-bit aggregation should be admitted, when execution should fall back to FP32, and which model regions are eligible for each route. The runtime preserves model semantics inside mixed DDP buckets and treats reducer admission as an architecture-runtime co-design problem rather than as a standalone compression operator. Across vision, Transformer, and autoregressive language-model workloads, NEURON-Fabric validates the path from calibration to distributed communication-hook execution. Static low-bit communication can collapse training accuracy, while profile-guided control preserves accuracy near full-precision references or calibrated targets and reduces modeled gradient-communication traffic in the evaluated settings. Transformer and billion-parameter language-model checks show that the same routing and fallback mechanisms execute across model families and multi-node deployments. Reducer-side replay and reducer-path measurements identify when compact sign-count aggregation is expected to reduce communication cost and when endpoint capacity should trigger fallback.
展开完整摘要收起摘要↓
Large-scale neural-network training repeatedly aggregates gradients across devices, making communication a central cost in distributed learning. Low-bit gradient aggregation can reduce this cost, but applying it as a static replacement for full-precision communication can destabilize training because safe precision depends on training phase, model structure, runtime bucketization, and the communication substrate. This paper presents NEURON-Fabric, a profile-guided runtime system for controlled low-bit gradient communication. NEURON-Fabric uses calibrated operating profiles, model-aware runtime bindings, online training-health monitoring, and reducer-capacity checks to decide when low-bit aggregation should be admitted, when execution should fall back to FP32, and which model regions are eligible for each route. The runtime preserves model semantics inside mixed DDP buckets and treats reducer admission as an architecture-runtime co-design problem rather than as a standalone compression operator. Across vision, Transformer, and autoregressive language-model workloads, NEURON-Fabric validates the path from calibration to distributed communication-hook execution. Static low-bit communication can collapse training accuracy, while profile-guided control preserves accuracy near full-precision references or calibrated targets and reduces modeled gradient-communication traffic in the evaluated settings. Transformer and billion-parameter language-model checks show that the same routing and fallback mechanisms execute across model families and multi-node deployments. Reducer-side replay and reducer-path measurements identify when compact sign-count aggregation is expected to reduce communication cost and when endpoint capacity should trigger fallback.
作者Wanning Zhang, Tongzhou Gu, Marco Canini, Ceyu Xu, Jian Weng
Large language model (LLM) inference is increasingly dominated by data movement across the memory hierarchy. Recent 3D-stacked cache technologies have enabled GB-scale last-level caches in modern server CPUs, making it possible to keep reusable model weights on chip and exploit cache bandwidth and latency. Achieving this regime is not straightforward: deeper pipelining for weight residency increases in-flight requests and KV-cache footprint, while cache-resident operators make operator-boundary synchronization a visible bottleneck. We present a cache-resident execution model for inference on hierarchical-memory clustered systems. The model separates weight-centric operators from attention and KV-cache management into dedicated resource domains, keeping reusable weights cache-resident while scaling KV capacity independently of pipeline depth. It also relaxes synchronization from operator boundaries to true sub-operator dependencies, reducing coordination overhead in the cache-resident regime. We instantiate this model on a multi-socket CPU cluster with a weight-attention decoupled architecture, locality-aware placement, and a specialized static runtime. The prototype substantially outperforms equally provisioned llama.cpp. On deployed Llama-3.2-3B and Llama-2-7B configurations, it achieves 2.04x-11.51x speedup on time-per-output-token (TPOT). Under a validated analytical model, it further reaches up to 13.9x TPOT speedup across model sizes, context lengths, and batch sizes. These results show that commodity CPUs with GB-scale last-level caches can support efficient LLM inference when execution is organized around cache residency, decoupled state management, and dependency-aware coordination.
展开完整摘要收起摘要↓
Large language model (LLM) inference is increasingly dominated by data movement across the memory hierarchy. Recent 3D-stacked cache technologies have enabled GB-scale last-level caches in modern server CPUs, making it possible to keep reusable model weights on chip and exploit cache bandwidth and latency. Achieving this regime is not straightforward: deeper pipelining for weight residency increases in-flight requests and KV-cache footprint, while cache-resident operators make operator-boundary synchronization a visible bottleneck. We present a cache-resident execution model for inference on hierarchical-memory clustered systems. The model separates weight-centric operators from attention and KV-cache management into dedicated resource domains, keeping reusable weights cache-resident while scaling KV capacity independently of pipeline depth. It also relaxes synchronization from operator boundaries to true sub-operator dependencies, reducing coordination overhead in the cache-resident regime. We instantiate this model on a multi-socket CPU cluster with a weight-attention decoupled architecture, locality-aware placement, and a specialized static runtime. The prototype substantially outperforms equally provisioned llama.cpp. On deployed Llama-3.2-3B and Llama-2-7B configurations, it achieves 2.04x-11.51x speedup on time-per-output-token (TPOT). Under a validated analytical model, it further reaches up to 13.9x TPOT speedup across model sizes, context lengths, and batch sizes. These results show that commodity CPUs with GB-scale last-level caches can support efficient LLM inference when execution is organized around cache residency, decoupled state management, and dependency-aware coordination.
作者Hualong Zhang, Siyang Feng, Zihan Huan, Yi Qian, Zhenbing Liu, Rushi Lan, Xipeng Pan
Histopathological tissue segmentation is essential for computer-aided diagnosis, yet weakly supervised methods often suffer from noisy pseudo-labels generated by Class Activation Mapping (CAM). Existing CAM approaches tend to focus on staining-driven appearance cues rather than true causal tissue morphology, resulting in spurious localization and poor structural consistency. To address this issue, we propose C$^2$RM-Seg, a two-stage framework that integrates causal pseudo-label refinement with structure-aware semantic enhancement. For classification, we introduce a Causal Counterfactual Reasoning Module (C$^2$RM) that decomposes features into latent factors and performs counterfactual intervention via a learned causal structure matrix, suppressing confounding context and producing morphology-aligned CAMs. For segmentation, we design a Dual-Path Structural-Semantic Architecture that combines fine-grained structural features from ResNeSt with global semantic priors from a frozen DINOV3 foundation model. A cross-path gating mechanism adaptively regulates semantic injection using local structural cues to preserve boundary fidelity. To further mitigate residual pseudo-label noise, we propose an Uncertainty-Gated Margin (UGM) loss, which dynamically balances margin enforcement and confidence learning based on prediction uncertainty. Extensive experiments on two public histopathological tissue datasets show that C$^2$RM-Seg achieves state-of-the-art performance.
展开完整摘要收起摘要↓
Histopathological tissue segmentation is essential for computer-aided diagnosis, yet weakly supervised methods often suffer from noisy pseudo-labels generated by Class Activation Mapping (CAM). Existing CAM approaches tend to focus on staining-driven appearance cues rather than true causal tissue morphology, resulting in spurious localization and poor structural consistency. To address this issue, we propose C$^2$RM-Seg, a two-stage framework that integrates causal pseudo-label refinement with structure-aware semantic enhancement. For classification, we introduce a Causal Counterfactual Reasoning Module (C$^2$RM) that decomposes features into latent factors and performs counterfactual intervention via a learned causal structure matrix, suppressing confounding context and producing morphology-aligned CAMs. For segmentation, we design a Dual-Path Structural-Semantic Architecture that combines fine-grained structural features from ResNeSt with global semantic priors from a frozen DINOV3 foundation model. A cross-path gating mechanism adaptively regulates semantic injection using local structural cues to preserve boundary fidelity. To further mitigate residual pseudo-label noise, we propose an Uncertainty-Gated Margin (UGM) loss, which dynamically balances margin enforcement and confidence learning based on prediction uncertainty. Extensive experiments on two public histopathological tissue datasets show that C$^2$RM-Seg achieves state-of-the-art performance.
With advances in autonomous driving and electric vehicle technologies, functional safety has become a critical requirement in automotive chip design. Traditional simulation-based fault analysis is often overly conservative at the module level and fails to accurately reflect fault criticality. This paper presents SafeGen, an LLM-driven, formal-verification-assisted framework for functional-safety-oriented fault criticality assessment. SafeGen leverages large language models (LLMs) and a document-level Hyper Knowledge Graph (HyperKG) that incorporates Failure Modes, Effects, and Diagnostic Analysis (FMEDA) guidelines to extract verifiable specifications from design and safety documents and evaluate their relevance to overall system safety. The HyperKG is further enriched with register-transfer-level (RTL) information to guide the generation of Functional Safety Assertions (FSAs) that are both semantically grounded and design-aware. Each assertion is linked to its corresponding specification, enabling traceable reasoning throughout the assessment process. A gate-to-RTL fault-mapping mechanism supporting both stuck-at and bridging faults, combined with formal property verification (FPV), enables semantic-level fault criticality grading based on specification-linked assertion violations. A digital-physical co-simulation platform for a field-oriented control (FOC) system is developed to validate SafeGen. Experimental results demonstrate that SafeGen generates higher-quality assertions than existing LLM-based assertion generation frameworks while providing greater semantic interpretability in fault criticality assessment compared with traditional simulation-based approaches.
展开完整摘要收起摘要↓
With advances in autonomous driving and electric vehicle technologies, functional safety has become a critical requirement in automotive chip design. Traditional simulation-based fault analysis is often overly conservative at the module level and fails to accurately reflect fault criticality. This paper presents SafeGen, an LLM-driven, formal-verification-assisted framework for functional-safety-oriented fault criticality assessment. SafeGen leverages large language models (LLMs) and a document-level Hyper Knowledge Graph (HyperKG) that incorporates Failure Modes, Effects, and Diagnostic Analysis (FMEDA) guidelines to extract verifiable specifications from design and safety documents and evaluate their relevance to overall system safety. The HyperKG is further enriched with register-transfer-level (RTL) information to guide the generation of Functional Safety Assertions (FSAs) that are both semantically grounded and design-aware. Each assertion is linked to its corresponding specification, enabling traceable reasoning throughout the assessment process. A gate-to-RTL fault-mapping mechanism supporting both stuck-at and bridging faults, combined with formal property verification (FPV), enables semantic-level fault criticality grading based on specification-linked assertion violations. A digital-physical co-simulation platform for a field-oriented control (FOC) system is developed to validate SafeGen. Experimental results demonstrate that SafeGen generates higher-quality assertions than existing LLM-based assertion generation frameworks while providing greater semantic interpretability in fault criticality assessment compared with traditional simulation-based approaches.
A language model's memory can be worse than having no memory at all. Give a model a memory that kept a wrong conclusion but dropped the work behind it, and it emits that stale value as a confident answer; give the same model an empty memory and it abstains. Across seven models this direction never reverses, a clean kill condition that none breaks. We call this brittle memory: behavioral, not the near-immediate information bound beneath it; only its magnitude is disposition- and task-dependent, not its direction. We measure it with reclaim evaluation: compress a drifted interaction at a fixed budget, then test whether a correction recovers the known answer, scored against ground truth with no judge. Correctability is bottlenecked by whether the answer-determining source survives, not by capability. A one-line source-first policy (keep the recomputable source, drop the re-derivable conclusion) restores correctability at equal budget where that source is compact and identifiable; a length-matched control rules out added text as the cause. The hand-built oracle reaches 1.00; a one-prompt deployable version reclaims 0.49-0.88. The stake compounds: chained through a memory loop, a single dropped-source error corrupts a growing span of downstream steps and stays uncorrectable, while source-first holds to a bounded budget horizon. The wall and fix replicate across three deployed memory systems and on real dialogue (MultiWOZ), and past the budget where the source no longer fits, the fix fails silently unless the note records completeness. This is a controlled study of a mechanism, not a benchmark: judge-free exact scoring, matched-budget controls, and validators built to come out false. We release the harness, conditions, and validators.
展开完整摘要收起摘要↓
A language model's memory can be worse than having no memory at all. Give a model a memory that kept a wrong conclusion but dropped the work behind it, and it emits that stale value as a confident answer; give the same model an empty memory and it abstains. Across seven models this direction never reverses, a clean kill condition that none breaks. We call this brittle memory: behavioral, not the near-immediate information bound beneath it; only its magnitude is disposition- and task-dependent, not its direction. We measure it with reclaim evaluation: compress a drifted interaction at a fixed budget, then test whether a correction recovers the known answer, scored against ground truth with no judge. Correctability is bottlenecked by whether the answer-determining source survives, not by capability. A one-line source-first policy (keep the recomputable source, drop the re-derivable conclusion) restores correctability at equal budget where that source is compact and identifiable; a length-matched control rules out added text as the cause. The hand-built oracle reaches 1.00; a one-prompt deployable version reclaims 0.49-0.88. The stake compounds: chained through a memory loop, a single dropped-source error corrupts a growing span of downstream steps and stays uncorrectable, while source-first holds to a bounded budget horizon. The wall and fix replicate across three deployed memory systems and on real dialogue (MultiWOZ), and past the budget where the source no longer fits, the fix fails silently unless the note records completeness. This is a controlled study of a mechanism, not a benchmark: judge-free exact scoring, matched-budget controls, and validators built to come out false. We release the harness, conditions, and validators.
Quantitative maps from dynamic contrast-enhanced MRI (DCE-MRI) are essential for tumor assessment but are often unavailable due to contrast-agent risks and protocol variability. Prior methods predict these maps from other MRI modalities, yet most assume fixed, fully observed inputs and fail under realistic missingness. We present Spatio-Temporal Mixture-of-Modality-Experts (ST-MoME), a conditional diffusion framework that synthesizes 3D DCE parameter maps from diverse subsets of multimodal MRI. ST-MoME fuses modality-specific expert features through a spatio-temporal gating network that produces voxel-wise, timestep-dependent weights, forming a conditioning tensor that guides denoising. To preserve quantitative fidelity, ST-MoME performs diffusion directly in image space with 3D patch-based training and a Swin-based backbone. On a clinical brain-tumor cohort of 386 patients, we evaluate ST-MoME across 16 controlled modality-availability scenarios. It achieves the lowest mean Normalized Mean Square Error (NMSE) aggregated across all three DCE parameters, with leading performance on $v_p$ and $v_e$, competitive results on $K^{\mathrm{trans}}$, and the lowest reconstruction error within the clinically critical tumor region. A post-hoc analysis of the learned gating dynamics shows a structural-early, physiological-late fusion schedule consistent with clinical intuition.
展开完整摘要收起摘要↓
Quantitative maps from dynamic contrast-enhanced MRI (DCE-MRI) are essential for tumor assessment but are often unavailable due to contrast-agent risks and protocol variability. Prior methods predict these maps from other MRI modalities, yet most assume fixed, fully observed inputs and fail under realistic missingness. We present Spatio-Temporal Mixture-of-Modality-Experts (ST-MoME), a conditional diffusion framework that synthesizes 3D DCE parameter maps from diverse subsets of multimodal MRI. ST-MoME fuses modality-specific expert features through a spatio-temporal gating network that produces voxel-wise, timestep-dependent weights, forming a conditioning tensor that guides denoising. To preserve quantitative fidelity, ST-MoME performs diffusion directly in image space with 3D patch-based training and a Swin-based backbone. On a clinical brain-tumor cohort of 386 patients, we evaluate ST-MoME across 16 controlled modality-availability scenarios. It achieves the lowest mean Normalized Mean Square Error (NMSE) aggregated across all three DCE parameters, with leading performance on $v_p$ and $v_e$, competitive results on $K^{\mathrm{trans}}$, and the lowest reconstruction error within the clinically critical tumor region. A post-hoc analysis of the learned gating dynamics shows a structural-early, physiological-late fusion schedule consistent with clinical intuition.
作者Yumeng Liu, Dongqi Wang, Jiyu Yu, Yijun Fan, Rong Xiong, Yue Wang
High-level humanoid planners often output sparse task-space, low-rate trajectories, whereas whole-body controllers run at high frequency. This creates temporal asynchrony between the planning and execution, and structural incompleteness for full-body control. We propose an asynchronous upper body task-space tracking framework for humanoids. A student policy is initialized by teacher-student distillation, conditioned on the full cached future trajectory and an execution-time index, and trained with a sliding-window global reward to reduce frame drift without explicit frame estimation. For task-specific post-training, an MPC module completes sparse references into floating-base and upper-body guidance, while action- and FK level self-guidance constrain policy drift. Simulation and Unitree G1 hardware experiments show improved tracking under low update rates, stronger performance than synchronous and decoupled baselines, and safer adaptation to out-of-distribution motions.
展开完整摘要收起摘要↓
High-level humanoid planners often output sparse task-space, low-rate trajectories, whereas whole-body controllers run at high frequency. This creates temporal asynchrony between the planning and execution, and structural incompleteness for full-body control. We propose an asynchronous upper body task-space tracking framework for humanoids. A student policy is initialized by teacher-student distillation, conditioned on the full cached future trajectory and an execution-time index, and trained with a sliding-window global reward to reduce frame drift without explicit frame estimation. For task-specific post-training, an MPC module completes sparse references into floating-base and upper-body guidance, while action- and FK level self-guidance constrain policy drift. Simulation and Unitree G1 hardware experiments show improved tracking under low update rates, stronger performance than synchronous and decoupled baselines, and safer adaptation to out-of-distribution motions.
作者Simon Kurgan, Evan Wang, Eric Leonen, Sophie Szeto, Luke Alexander, Artemii Remizov, Jarod Alper, Giovanni Inchiostro, Vasily Ilin
Mathematical knowledge is organized around statements and their dependencies, but this structure is exposed unevenly: informal papers cite mostly at the document level, while formal libraries record fine-grained dependencies over a much smaller body of mathematics. We introduce TheoremGraph, a unified statement-level dependency graph spanning both informal and formal mathematics. On the informal side, we parse 11.7M theorem-like environments from mathematics arXiv and recover 18.3M candidate directed dependencies, each labeled by the extractor that proposed it so downstream users can trade coverage for precision. On the formal side, we release LeanGraph, a Lean 4 elaborator-level extractor producing 388,105 declaration nodes and 11.3M typed edges across 25 Lean projects. We bridge the two graphs by embedding generated natural-language slogans into a shared semantic space, linking related statements across papers and across the informal/formal divide; an LLM judge affirms 47,952 such matches above a 0.8 cosine floor, with the judge-acceptance rate rising from 48% across the floor to 87% in the >=0.9 tier. On formal concept retrieval, our name-and-signature representation with graph expansion comes within 0.5pp of LeanSearch v2's reranked Recall@10 (0.775 vs. 0.780) without an LM reranker. We release the dataset, extractors, HTTP API, and MCP interface as infrastructure for mathematical search, attribution, and retrieval-augmented reasoning, available at theoremsearch.com and huggingface.co/datasets/uw-math-ai/theorem-matching.
展开完整摘要收起摘要↓
Mathematical knowledge is organized around statements and their dependencies, but this structure is exposed unevenly: informal papers cite mostly at the document level, while formal libraries record fine-grained dependencies over a much smaller body of mathematics. We introduce TheoremGraph, a unified statement-level dependency graph spanning both informal and formal mathematics. On the informal side, we parse 11.7M theorem-like environments from mathematics arXiv and recover 18.3M candidate directed dependencies, each labeled by the extractor that proposed it so downstream users can trade coverage for precision. On the formal side, we release LeanGraph, a Lean 4 elaborator-level extractor producing 388,105 declaration nodes and 11.3M typed edges across 25 Lean projects. We bridge the two graphs by embedding generated natural-language slogans into a shared semantic space, linking related statements across papers and across the informal/formal divide; an LLM judge affirms 47,952 such matches above a 0.8 cosine floor, with the judge-acceptance rate rising from 48% across the floor to 87% in the >=0.9 tier. On formal concept retrieval, our name-and-signature representation with graph expansion comes within 0.5pp of LeanSearch v2's reranked Recall@10 (0.775 vs. 0.780) without an LM reranker. We release the dataset, extractors, HTTP API, and MCP interface as infrastructure for mathematical search, attribution, and retrieval-augmented reasoning, available at theoremsearch.com and huggingface.co/datasets/uw-math-ai/theorem-matching.
While large language model (LLM)-based text-to-speech (TTS) systems have achieved high-quality speech synthesis, most existing systems focus on English and Chinese. Japanese, however, remains under-explored, and its unique linguistic challenges, such as widespread context-dependent kanji polyphony, have yet to be adequately tackled. Here we introduce Sarashina2.2-TTS (https://github.com/sbintuitions/sarashina2.2-tts), a Japanese-centric LLM-TTS system that tackles these challenges through a dual approach: data strategy and evaluation methodology. First, we scale training to approximately 361k hours of speech, incorporating a balanced mix of Japanese and English data. Furthermore, we design a targeted data augmentation pipeline covering all 2,136 Joyo (regular-use) kanji designated by Japan's Agency for Cultural Affairs to efficiently address kanji polyphony disambiguation. Second, we introduce the Joyo Kanji Yomi Benchmark (https://github.com/sbintuitions/JoyoKanji-Yomi-Benchmark), covering all 2,136 Joyo kanji and their 4,378 readings. Alongside this benchmark, we propose Kana-CER, a metric that compares synthesized speech against reference readings in the kana space, eliminating orthographic variations to directly measure pronunciation correctness. Experiments demonstrate that our targeted data augmentation significantly improves reading accuracy. Overall, Sarashina2.2-TTS achieves state-of-the-art kanji-level reading accuracy and matches top baselines on general sentence-level pronunciation, while delivering the highest speaker similarity in zero-shot Japanese speech synthesis. Furthermore, cross-lingual evaluation reveals that Sarashina2.2-TTS is the only system that maintains stable Japanese pronunciation regardless of the prompt language, confirming that our balanced training approach improves cross-lingual robustness.
展开完整摘要收起摘要↓
While large language model (LLM)-based text-to-speech (TTS) systems have achieved high-quality speech synthesis, most existing systems focus on English and Chinese. Japanese, however, remains under-explored, and its unique linguistic challenges, such as widespread context-dependent kanji polyphony, have yet to be adequately tackled. Here we introduce Sarashina2.2-TTS (https://github.com/sbintuitions/sarashina2.2-tts), a Japanese-centric LLM-TTS system that tackles these challenges through a dual approach: data strategy and evaluation methodology. First, we scale training to approximately 361k hours of speech, incorporating a balanced mix of Japanese and English data. Furthermore, we design a targeted data augmentation pipeline covering all 2,136 Joyo (regular-use) kanji designated by Japan's Agency for Cultural Affairs to efficiently address kanji polyphony disambiguation. Second, we introduce the Joyo Kanji Yomi Benchmark (https://github.com/sbintuitions/JoyoKanji-Yomi-Benchmark), covering all 2,136 Joyo kanji and their 4,378 readings. Alongside this benchmark, we propose Kana-CER, a metric that compares synthesized speech against reference readings in the kana space, eliminating orthographic variations to directly measure pronunciation correctness. Experiments demonstrate that our targeted data augmentation significantly improves reading accuracy. Overall, Sarashina2.2-TTS achieves state-of-the-art kanji-level reading accuracy and matches top baselines on general sentence-level pronunciation, while delivering the highest speaker similarity in zero-shot Japanese speech synthesis. Furthermore, cross-lingual evaluation reveals that Sarashina2.2-TTS is the only system that maintains stable Japanese pronunciation regardless of the prompt language, confirming that our balanced training approach improves cross-lingual robustness.
作者Ching-Hao Chiu, Hao-Wei Chung, Gelei Xu, Xueyang Li, Pin-Yu Chen, John Kheir, Meysam Ghaffari, Carlos Morato, Ahmed Abbasi, Yiyu Shi
With the rapid adoption of generative AI, synthetic medical images pose growing risks, including diagnostic deception and insurance fraud. Although prior work has explored vision-language model (VLM)-based synthetic image detection, these evaluations typically consider images in isolation. In clinical practice, however, images are interpreted alongside structured records and metadata, and VLMs are increasingly deployed under joint image-record inputs. We uncover a previously underexamined multimodal vulnerability: when given both modalities, VLMs may overweight record context in authenticity judgments, such that the same image receives different predictions solely due to changes in its accompanying text. This raises concerns about robustness in real-world deployment. To systematically characterize this effect, we reformulate synthetic medical image detection as an audit of multimodal robustness at the image-record interface and introduce a paired benchmark that holds the image fixed while swapping controlled metadata variants. Across multiple imaging modalities, we evaluate diverse open-weight and frontier API VLMs and quantify how metadata alone shifts authenticity predictions. Our benchmark provides a standardized tool for assessing and improving multimodal robustness beyond image-only settings. The code is available at https://github.com/chiuhaohao/Beyond-Visual-Forensics.
展开完整摘要收起摘要↓
With the rapid adoption of generative AI, synthetic medical images pose growing risks, including diagnostic deception and insurance fraud. Although prior work has explored vision-language model (VLM)-based synthetic image detection, these evaluations typically consider images in isolation. In clinical practice, however, images are interpreted alongside structured records and metadata, and VLMs are increasingly deployed under joint image-record inputs. We uncover a previously underexamined multimodal vulnerability: when given both modalities, VLMs may overweight record context in authenticity judgments, such that the same image receives different predictions solely due to changes in its accompanying text. This raises concerns about robustness in real-world deployment. To systematically characterize this effect, we reformulate synthetic medical image detection as an audit of multimodal robustness at the image-record interface and introduce a paired benchmark that holds the image fixed while swapping controlled metadata variants. Across multiple imaging modalities, we evaluate diverse open-weight and frontier API VLMs and quantify how metadata alone shifts authenticity predictions. Our benchmark provides a standardized tool for assessing and improving multimodal robustness beyond image-only settings. The code is available at https://github.com/chiuhaohao/Beyond-Visual-Forensics.
作者Sitong Cheng, Weizhen Bian, Songjun Cao, Jin Li, Bei Liu, Chunyang Jiang, Yike Zhang, Weihao Wu, Yiming Li, Chi-Min Chan, Long Ma, Wei Xue
Speech-to-speech translation (S2ST) should preserve not only lexical meaning, but also expressive attributes: emotion, scenario style (e.g., news reporting vs. dramatic dialogue), and nonverbal vocalizations (NVs). Moreover, collecting cross-lingual target speech that is both translation-faithful and expressively aligned with the source is difficult at scale, making reference-based evaluation impractical. We introduce STEB (Speech-to-Speech Translation Expressiveness Benchmark), a 32.6-hour Chinese--English benchmark that evaluates both standard dimensions (translation fidelity, speaker similarity, duration alignment) and expressiveness dimensions (emotion, scenario style, NV preservation). For expressiveness evaluation, STEB uses a caption-then-summarize framework that converts speech into structured expressive attributes and compares source and hypothesis attributes with an LLM judge. Human validation shows statistically significant correlations with listener judgments across all expressive dimensions. We evaluate six S2ST systems covering cascaded systems, end-to-end models, and speech large language models. Many systems, especially cascaded ones, achieve strong translation fidelity, but they still struggle with emotion preservation (best: 3.82/5) and NV preservation (best: 2.31/5). These results reveal a gap between semantic transfer and expressive transfer, identifying expressiveness preservation as an open challenge for S2ST. Audio samples are available at https://cmots.github.io/steb.github.io/.
展开完整摘要收起摘要↓
Speech-to-speech translation (S2ST) should preserve not only lexical meaning, but also expressive attributes: emotion, scenario style (e.g., news reporting vs. dramatic dialogue), and nonverbal vocalizations (NVs). Moreover, collecting cross-lingual target speech that is both translation-faithful and expressively aligned with the source is difficult at scale, making reference-based evaluation impractical. We introduce STEB (Speech-to-Speech Translation Expressiveness Benchmark), a 32.6-hour Chinese--English benchmark that evaluates both standard dimensions (translation fidelity, speaker similarity, duration alignment) and expressiveness dimensions (emotion, scenario style, NV preservation). For expressiveness evaluation, STEB uses a caption-then-summarize framework that converts speech into structured expressive attributes and compares source and hypothesis attributes with an LLM judge. Human validation shows statistically significant correlations with listener judgments across all expressive dimensions. We evaluate six S2ST systems covering cascaded systems, end-to-end models, and speech large language models. Many systems, especially cascaded ones, achieve strong translation fidelity, but they still struggle with emotion preservation (best: 3.82/5) and NV preservation (best: 2.31/5). These results reveal a gap between semantic transfer and expressive transfer, identifying expressiveness preservation as an open challenge for S2ST. Audio samples are available at https://cmots.github.io/steb.github.io/.
作者Jiayu Li, Yixiao Fang, Tianyu Hu, Wei Cheng, Ping Huang, Zheheng Fan, Gang Yu, Xingjun Ma
Real-world photography requires capture-time guidance for both camera framing and subject pose. Yet existing aesthetic cropping benchmarks mainly evaluate post-hoc crop prediction and overlook subject-side recommendations, leaving the capture-time guidance capabilities of multimodal large language models (MLLMs) underexplored. To address this gap, we introduce CaptureGuide-Bench, a benchmark with two complementary tasks: photographer-side composition decision and refinement, and subject-side scene-conditioned pose recommendation. Our evaluation reveals limitations: general-purpose MLLMs can make composition decisions but lack precise refinement localization, while specialized aesthetic cropping models localize crops effectively but are limited to refinement; neither provides actionable pose guidance. To support model development, we further construct CaptureGuide-Dataset, comprising 130K samples with textual rationales and structured visual annotations, and develop ShutterMuse, a unified MLLM trained with supervised and reinforcement fine-tuning. Experiments on CaptureGuide-Bench show that ShutterMuse achieves the best overall photographer-side performance among evaluated baselines and competitive subject-side pose recommendation with substantially lower inference cost, demonstrating the potential of MLLMs as interactive assistants for photography during image capture.
展开完整摘要收起摘要↓
Real-world photography requires capture-time guidance for both camera framing and subject pose. Yet existing aesthetic cropping benchmarks mainly evaluate post-hoc crop prediction and overlook subject-side recommendations, leaving the capture-time guidance capabilities of multimodal large language models (MLLMs) underexplored. To address this gap, we introduce CaptureGuide-Bench, a benchmark with two complementary tasks: photographer-side composition decision and refinement, and subject-side scene-conditioned pose recommendation. Our evaluation reveals limitations: general-purpose MLLMs can make composition decisions but lack precise refinement localization, while specialized aesthetic cropping models localize crops effectively but are limited to refinement; neither provides actionable pose guidance. To support model development, we further construct CaptureGuide-Dataset, comprising 130K samples with textual rationales and structured visual annotations, and develop ShutterMuse, a unified MLLM trained with supervised and reinforcement fine-tuning. Experiments on CaptureGuide-Bench show that ShutterMuse achieves the best overall photographer-side performance among evaluated baselines and competitive subject-side pose recommendation with substantially lower inference cost, demonstrating the potential of MLLMs as interactive assistants for photography during image capture.
作者Shangkun Li, Jie Xu, Yi Guo, Zeju Li, Yuanyuan Wang
Medical vision-language models typically generate diagnoses through single-pass inference without indicating which image regions support their conclusions. This lack of spatial grounding limits clinical utility: outputs cannot be audited, and models may hallucinate findings on normal scans. We present BrReMark (Brain Rethink via ROI Marking), a framework that introduces explicit region marking into brain MRI diagnosis. The model first generates hypotheses about potential abnormalities and grounds them through explicit bounding box marking, then verifies conclusions by re-examining the marked evidence. Training combines supervised fine-tuning on structured reasoning trajectories with reinforcement learning using a composite reward over localization accuracy and diagnostic reasoning. Furthermore, we integrate a domain randomization-based pathology synthesis augmentation strategy to improve the model's generalizability to out-of-distribution (OOD) data. On internal benchmark, BrReMark improves mAP50 from 0.74% to 37.54% compared to the base model, while achieving 21.57% Clinical F1 and 45.26% diagnostic accuracy. On NOVA OOD benchmark, it also achieves competitive overall performance with a 45.7% reduction in false positives compared to the state-of-the-art, indicating reduced hallucination on rare pathologies. These findings suggest that explicit hypothesis-verification grounding is a practical path toward trustworthy open-ended brain MRI diagnosis across both in-distribution and OOD settings.
展开完整摘要收起摘要↓
Medical vision-language models typically generate diagnoses through single-pass inference without indicating which image regions support their conclusions. This lack of spatial grounding limits clinical utility: outputs cannot be audited, and models may hallucinate findings on normal scans. We present BrReMark (Brain Rethink via ROI Marking), a framework that introduces explicit region marking into brain MRI diagnosis. The model first generates hypotheses about potential abnormalities and grounds them through explicit bounding box marking, then verifies conclusions by re-examining the marked evidence. Training combines supervised fine-tuning on structured reasoning trajectories with reinforcement learning using a composite reward over localization accuracy and diagnostic reasoning. Furthermore, we integrate a domain randomization-based pathology synthesis augmentation strategy to improve the model's generalizability to out-of-distribution (OOD) data. On internal benchmark, BrReMark improves mAP50 from 0.74% to 37.54% compared to the base model, while achieving 21.57% Clinical F1 and 45.26% diagnostic accuracy. On NOVA OOD benchmark, it also achieves competitive overall performance with a 45.7% reduction in false positives compared to the state-of-the-art, indicating reduced hallucination on rare pathologies. These findings suggest that explicit hypothesis-verification grounding is a practical path toward trustworthy open-ended brain MRI diagnosis across both in-distribution and OOD settings.
Large language models are turning from isolated predictors into agentic systems: they call tools, retrieve evidence, obey environment constraints, use verifiers, and complete tasks through search and multi-turn interaction. We adopts an analytical viewpoint based on "compression is intelligence": under a fixed task distribution, interface, and compute budget, a stronger agentic system lets a target object be reconstructed with fewer bits. We operationalize the measure with arithmetic coding, seed coding, and a fallback, and evaluate it in five settings: reversed text, chess moves, protein sequences, retrieval-augmented question answering, and semantic story compression; in all of them agentic components reduce codelength. These small, controlled experiments cover component types typical of real agentic systems, show that codelength can analyze how components, observers, and budgets change residual uncertainty, and offer guidance for evaluating real agent systems.
展开完整摘要收起摘要↓
Large language models are turning from isolated predictors into agentic systems: they call tools, retrieve evidence, obey environment constraints, use verifiers, and complete tasks through search and multi-turn interaction. We adopts an analytical viewpoint based on "compression is intelligence": under a fixed task distribution, interface, and compute budget, a stronger agentic system lets a target object be reconstructed with fewer bits. We operationalize the measure with arithmetic coding, seed coding, and a fallback, and evaluate it in five settings: reversed text, chess moves, protein sequences, retrieval-augmented question answering, and semantic story compression; in all of them agentic components reduce codelength. These small, controlled experiments cover component types typical of real agentic systems, show that codelength can analyze how components, observers, and budgets change residual uncertainty, and offer guidance for evaluating real agent systems.
作者Ilia Kulikov, Chenxi Whitehouse, Tianhao Wu, Yixin Nie, Swarnadeep Saha, Eryk Helenowski, Weizhe Yuan, Olga Golovneva, Jack Lanchantin, Yoram Bachrach, Jakob Foerster, Xian Li, Han Fang, Sainbayar Sukhbaatar, Jason Weston
We introduce Autodata, a general method that enables AI agents to act as data scientists who build high quality training and evaluation data. We show how to train (meta-optimize) such a data scientist agent, so that it learns to create even stronger data. We describe the overall formulation, and a specific practical implementation, Agentic Self-Instruct. We conduct experiments on computer science research tasks, legal reasoning tasks and reasoning with mathematical objects, where we obtain improved results compared to classical synthetic dataset creation methods. Further, meta-optimizing the data scientist agent itself delivers an even larger performance uplift. Agentic data creation provides a way to convert increased inference compute into higher quality model training. Overall, we believe this direction has the potential to change the way we build AI data.
展开完整摘要收起摘要↓
We introduce Autodata, a general method that enables AI agents to act as data scientists who build high quality training and evaluation data. We show how to train (meta-optimize) such a data scientist agent, so that it learns to create even stronger data. We describe the overall formulation, and a specific practical implementation, Agentic Self-Instruct. We conduct experiments on computer science research tasks, legal reasoning tasks and reasoning with mathematical objects, where we obtain improved results compared to classical synthetic dataset creation methods. Further, meta-optimizing the data scientist agent itself delivers an even larger performance uplift. Agentic data creation provides a way to convert increased inference compute into higher quality model training. Overall, we believe this direction has the potential to change the way we build AI data.
Large language models (LLMs) have shown promising performance across a wide range of biomedical applications, including medical question answering (QA), yet they remain prone to hallucinations and outdated knowledge. Although retrieval-augmented generation (RAG) can alleviate this issue by incorporating external documents, there still exist two fundamental limitations. First, medical knowledge is often fragmented across documents, while most RAG methods rely on a single retrieval path, which makes it challenging to jointly preserve fine-grained semantic information and structured global associations. Second, static retrieval strategies are typically insufficient to support deep reasoning that is important in complex medical QA. In this paper, we present a dual-path retrieval framework with an iterative retrieval-reasoning mechanism termed "Hybrid-IR" for complex medical QA. The proposed Hybrid-IR integrates graph-based retrieval for exploration of structured knowledge and dense retrieval for fine-grained semantic matching. Moreover, the reasoning trajectory can be progressively refined through an iterative retrieve-reason loop. Experiments on three widely used medical QA benchmarks demonstrate the effectiveness of our Hybrid-IR.
展开完整摘要收起摘要↓
Large language models (LLMs) have shown promising performance across a wide range of biomedical applications, including medical question answering (QA), yet they remain prone to hallucinations and outdated knowledge. Although retrieval-augmented generation (RAG) can alleviate this issue by incorporating external documents, there still exist two fundamental limitations. First, medical knowledge is often fragmented across documents, while most RAG methods rely on a single retrieval path, which makes it challenging to jointly preserve fine-grained semantic information and structured global associations. Second, static retrieval strategies are typically insufficient to support deep reasoning that is important in complex medical QA. In this paper, we present a dual-path retrieval framework with an iterative retrieval-reasoning mechanism termed "Hybrid-IR" for complex medical QA. The proposed Hybrid-IR integrates graph-based retrieval for exploration of structured knowledge and dense retrieval for fine-grained semantic matching. Moreover, the reasoning trajectory can be progressively refined through an iterative retrieve-reason loop. Experiments on three widely used medical QA benchmarks demonstrate the effectiveness of our Hybrid-IR.
Large language models (LLMs) are increasingly deployed across languages, but their safety behavior remains uneven across linguistic and cultural contexts. This survey synthesizes work on toxicity detection and detoxification for multilingual LLMs. We first catalogue threat models that exploit language choice, translation pivots, code-switching, orthographic variation, multi-turn interaction, and post-deployment fine-tuning to weaken safety alignment. We then organize task formulations (toxic-to-neutral rewriting, toxicity classification, and toxic-generation evaluation), multilingual detection approaches (cross-lingual encoders, translation pipelines, representation-level probes, and LLM-based detectors), and mitigation strategies spanning data filtering, supervised and preference-based tuning, decoding-time steering, representation editing, and multilingual guardrails. Across these areas, we identify persistent challenges: uneven language coverage, culturally contingent definitions of harm, fragmented evaluation protocols, and the risk that detoxification suppresses legitimate dialectal or identity-related expression.
展开完整摘要收起摘要↓
Large language models (LLMs) are increasingly deployed across languages, but their safety behavior remains uneven across linguistic and cultural contexts. This survey synthesizes work on toxicity detection and detoxification for multilingual LLMs. We first catalogue threat models that exploit language choice, translation pivots, code-switching, orthographic variation, multi-turn interaction, and post-deployment fine-tuning to weaken safety alignment. We then organize task formulations (toxic-to-neutral rewriting, toxicity classification, and toxic-generation evaluation), multilingual detection approaches (cross-lingual encoders, translation pipelines, representation-level probes, and LLM-based detectors), and mitigation strategies spanning data filtering, supervised and preference-based tuning, decoding-time steering, representation editing, and multilingual guardrails. Across these areas, we identify persistent challenges: uneven language coverage, culturally contingent definitions of harm, fragmented evaluation protocols, and the risk that detoxification suppresses legitimate dialectal or identity-related expression.
作者Yuchen Xiao, Lei Yuan, Ruiqi Xue, Tieyue Yin, Yang Yu
Extracting skills from multi-agent offline dataset improves learning efficiency via sharing task-invariant coordination skills among tasks. In settings where tasks occur sequentially and the space of skills grows exponentially, existing approaches that rely on heuristically designed and fixed-sized skill libraries struggle to resolve the problem of distributional shift and interference, facing catastrophic forgetting and plasticity loss. To address this problem and endow agents with the ability to continually discover and reuse coordination skills in open-environment, we propose COMAD, a principled framework for Continual Offline Multi-agent Skill Discovery via Skill Partition and Reuse. We first discover skills from mixed multi-agent behavior data with an auto-encoder to transform coordination knowledge into reusable coordination skills. Then we construct a skill-augmented policy learning objective with multi-head architectures, explicitly guiding the advantage function with reusable skills identified via a density-based reusability estimator. Theoretical analysis shows our method approximates the optimum of a continual skill discovery problem. Empirical results across diverse MARL benchmarks show that COMAD continually expands its skill library to mitigate interference, achieving superior forward and backward transfer for task streams compared to multiple baselines.
展开完整摘要收起摘要↓
Extracting skills from multi-agent offline dataset improves learning efficiency via sharing task-invariant coordination skills among tasks. In settings where tasks occur sequentially and the space of skills grows exponentially, existing approaches that rely on heuristically designed and fixed-sized skill libraries struggle to resolve the problem of distributional shift and interference, facing catastrophic forgetting and plasticity loss. To address this problem and endow agents with the ability to continually discover and reuse coordination skills in open-environment, we propose COMAD, a principled framework for Continual Offline Multi-agent Skill Discovery via Skill Partition and Reuse. We first discover skills from mixed multi-agent behavior data with an auto-encoder to transform coordination knowledge into reusable coordination skills. Then we construct a skill-augmented policy learning objective with multi-head architectures, explicitly guiding the advantage function with reusable skills identified via a density-based reusability estimator. Theoretical analysis shows our method approximates the optimum of a continual skill discovery problem. Empirical results across diverse MARL benchmarks show that COMAD continually expands its skill library to mitigate interference, achieving superior forward and backward transfer for task streams compared to multiple baselines.
With the widespread adoption of large language models (LLMs) in chatbots and everyday applications, companies increasingly need guardrails that are effective while remaining low-cost and low-latency. Safety evaluation of LLM outputs has generally relied on LLM-based judges, which can be effective but are often slow and expensive to deploy at scale. In this paper, we evaluate whether fine-tuned modern encoder classifiers from the ModernBERT family, including ModernBERT and Ettin, can reliably identify harmful LLM outputs in user-model conversations without substantial performance loss relative to LLM-based judges. We benchmark these encoder classifiers against rule-based prefix matching, fine-tuned LLM classifiers, and LLM judges using a range of judge-prompting strategies across open-source adversarial datasets. The LLM judges include evaluation methodologies from StrongReject, ShieldGemma, JailbreakBench, AILuminate, SorryBench, and a Claude-as-a-judge setup, as well as fine-tuned safety classifiers such as LlamaGuard 3 and LlamaGuard 4. The encoder classifiers are fine-tuned on judge-labeled data using a majority-voting label strategy and are then evaluated on a gold-standard holdout dataset to assess their performance relative to LLM judges. We report absolute performance using F1 score, false negative rate, and precision-recall metrics. We also break down results by attack technique, including single-turn prompting, decomposition, escalation, and context manipulation, to identify where encoder classifiers align with or diverge from LLM-based judges. Our findings provide guidance on when encoder classifiers can serve as cost- and latency-efficient alternatives to LLM-based safety evaluation.
展开完整摘要收起摘要↓
With the widespread adoption of large language models (LLMs) in chatbots and everyday applications, companies increasingly need guardrails that are effective while remaining low-cost and low-latency. Safety evaluation of LLM outputs has generally relied on LLM-based judges, which can be effective but are often slow and expensive to deploy at scale. In this paper, we evaluate whether fine-tuned modern encoder classifiers from the ModernBERT family, including ModernBERT and Ettin, can reliably identify harmful LLM outputs in user-model conversations without substantial performance loss relative to LLM-based judges. We benchmark these encoder classifiers against rule-based prefix matching, fine-tuned LLM classifiers, and LLM judges using a range of judge-prompting strategies across open-source adversarial datasets. The LLM judges include evaluation methodologies from StrongReject, ShieldGemma, JailbreakBench, AILuminate, SorryBench, and a Claude-as-a-judge setup, as well as fine-tuned safety classifiers such as LlamaGuard 3 and LlamaGuard 4. The encoder classifiers are fine-tuned on judge-labeled data using a majority-voting label strategy and are then evaluated on a gold-standard holdout dataset to assess their performance relative to LLM judges. We report absolute performance using F1 score, false negative rate, and precision-recall metrics. We also break down results by attack technique, including single-turn prompting, decomposition, escalation, and context manipulation, to identify where encoder classifiers align with or diverge from LLM-based judges. Our findings provide guidance on when encoder classifiers can serve as cost- and latency-efficient alternatives to LLM-based safety evaluation.
Large language models (LLMs) are increasingly used to generate personalized guidance in wellbeing contexts such as physical activity, stress management, and mental health support, enabling fluent and context-aware interaction but relying on largely implicit mechanisms that shape how recommendations are expressed and adapted. We argue that this reliance on implicit adaptation through prompting and alignment limits control over guidance, responsibility framing, and user influence, which is particularly problematic in wellbeing settings where recommendations affect users' actions and long-term outcomes. We propose a system-level perspective in which conversational behavior is structured through explicit interaction constraints, including guidance strategies, explanation styles, degrees of directness, and mechanisms for user control. Building on prior work on tangible recommendations, we show how these constraints address key challenges in wellbeing-oriented recommendation, namely trust calibration, intent alignment, and consequence awareness. We outline a modular architecture for controllable LLM-based recommendation and discuss how different configurations can be systematically designed and evaluated in relation to user-centered outcomes such as self-efficacy, perceived agency, and appropriate reliance. This paper contributes a system-level framework for designing LLM-based recommender systems that are adaptive while remaining transparent, controllable, and aligned with human wellbeing.
展开完整摘要收起摘要↓
Large language models (LLMs) are increasingly used to generate personalized guidance in wellbeing contexts such as physical activity, stress management, and mental health support, enabling fluent and context-aware interaction but relying on largely implicit mechanisms that shape how recommendations are expressed and adapted. We argue that this reliance on implicit adaptation through prompting and alignment limits control over guidance, responsibility framing, and user influence, which is particularly problematic in wellbeing settings where recommendations affect users' actions and long-term outcomes. We propose a system-level perspective in which conversational behavior is structured through explicit interaction constraints, including guidance strategies, explanation styles, degrees of directness, and mechanisms for user control. Building on prior work on tangible recommendations, we show how these constraints address key challenges in wellbeing-oriented recommendation, namely trust calibration, intent alignment, and consequence awareness. We outline a modular architecture for controllable LLM-based recommendation and discuss how different configurations can be systematically designed and evaluated in relation to user-centered outcomes such as self-efficacy, perceived agency, and appropriate reliance. This paper contributes a system-level framework for designing LLM-based recommender systems that are adaptive while remaining transparent, controllable, and aligned with human wellbeing.
Large language models are increasingly deployed as investment research assistants, yet no benchmark tests whether they can accurately reconstruct and apply the specific procedural decision frameworks of expert investors. We introduce InvestPhilBench, a multi-layer dynamic benchmark spanning eight cognitive tiers, from principle identification (L1) to novel framework extrapolation (L8). The v0.6 release comprises 118 primary-source-verified investment principle cards, 25 decision framework cards with explicit topology metadata, and 243 QA questions (197 dev / 46 held-out test). For reproducible scoring at scale we introduce the Benchmark Automated Scoring Pipeline (BASP) — five algorithmic metrics (OGRS, KCCS, SAP@k, IVP, CKCA) — the Failure Mode Detection Protocol (FMDP) with computable rules for six failure modes, and Gate Reconstruction Accuracy (GRA), a per-gate metric for questions with gold reasoning programs. In this release, InvestPhilBench is primarily a benchmark-and-methodology contribution. A four-model sanity wave on the 188-question development split shows a sharp provider-tier split (BASP 0.906 vs. 0.438); these mixed-judge numbers are confounded upper bounds. The central finding: the BASP composite saturates at the frontier (Claude L4 = 0.932) while GRA still exposes a procedural deficit (frontier L4 GRA approx. 0.77, L7 GRA 0.57-0.62) — composite scoring rewards fluent prose and hides the procedural gap. v0.6 implements a unified judge and true model-in-the-loop retrieval/oracle conditions; the de-confounded multi-model leaderboard and full three-condition run are v1.0 deliverables. On a 100-item expert-annotated gold set the automated BASP composite tracks the human reference at Pearson r = 0.72 (MAE = 0.10), with attribution (SAP@3) the weakest sub-metric and the failure-mode detector running sensitive-but-over-flagging.
展开完整摘要收起摘要↓
Large language models are increasingly deployed as investment research assistants, yet no benchmark tests whether they can accurately reconstruct and apply the specific procedural decision frameworks of expert investors. We introduce InvestPhilBench, a multi-layer dynamic benchmark spanning eight cognitive tiers, from principle identification (L1) to novel framework extrapolation (L8). The v0.6 release comprises 118 primary-source-verified investment principle cards, 25 decision framework cards with explicit topology metadata, and 243 QA questions (197 dev / 46 held-out test). For reproducible scoring at scale we introduce the Benchmark Automated Scoring Pipeline (BASP) — five algorithmic metrics (OGRS, KCCS, SAP@k, IVP, CKCA) — the Failure Mode Detection Protocol (FMDP) with computable rules for six failure modes, and Gate Reconstruction Accuracy (GRA), a per-gate metric for questions with gold reasoning programs. In this release, InvestPhilBench is primarily a benchmark-and-methodology contribution. A four-model sanity wave on the 188-question development split shows a sharp provider-tier split (BASP 0.906 vs. 0.438); these mixed-judge numbers are confounded upper bounds. The central finding: the BASP composite saturates at the frontier (Claude L4 = 0.932) while GRA still exposes a procedural deficit (frontier L4 GRA approx. 0.77, L7 GRA 0.57-0.62) — composite scoring rewards fluent prose and hides the procedural gap. v0.6 implements a unified judge and true model-in-the-loop retrieval/oracle conditions; the de-confounded multi-model leaderboard and full three-condition run are v1.0 deliverables. On a 100-item expert-annotated gold set the automated BASP composite tracks the human reference at Pearson r = 0.72 (MAE = 0.10), with attribution (SAP@3) the weakest sub-metric and the failure-mode detector running sensitive-but-over-flagging.
作者Yves Ferstler, Adam Podoxin, Ty Brassington, Roman Grundkiewicz, Maite Taboada, Marzena Karpinska
AI translation of literary works is increasingly common. While the content may be rendered adequately, we do not know enough about how readers experience it in terms of immersiveness and literary effect, aspects poorly captured by automatic machine translation metrics or human evaluation targeting fluency and adequacy. We ask 15 avid readers to compare recently published human translations (HT) to machine translations (MT) generated with an agentic large language model (LLM)-based pipeline, for 15 recent novels in French, Polish, and Japanese and translated into English. Readers evaluated approximately 8K-word excerpts in two conditions: immersive reading of the whole excerpt (30 comparisons) and close reading of 386 aligned HT-MT chunk pairs (772 comparisons), with two readers per book and in alternating order of presentation. Overall, readers find MT "fine", but prefer HT (slightly at excerpt-level 19/30, more clearly at chunk-level 522/772) for its ease, clarity, and immersive nature. Readers' highlights show that MT's quality varies more within one book than HT's does. Crucially, readers cannot reliably tell the two apart (17/30 guess correctly) and tend to prefer the version they believe to be human. Automatic metrics, including LLM-as-a-judge approaches, fail to recover reader preferences and favor MT. We release LAIT (Literary AI Translation), a reader-centered evaluation dataset with 1K reader comments, 2K judgments and preference ratings, and 7.2K span-level annotations, along with our evaluation protocol and supporting interface.
展开完整摘要收起摘要↓
AI translation of literary works is increasingly common. While the content may be rendered adequately, we do not know enough about how readers experience it in terms of immersiveness and literary effect, aspects poorly captured by automatic machine translation metrics or human evaluation targeting fluency and adequacy. We ask 15 avid readers to compare recently published human translations (HT) to machine translations (MT) generated with an agentic large language model (LLM)-based pipeline, for 15 recent novels in French, Polish, and Japanese and translated into English. Readers evaluated approximately 8K-word excerpts in two conditions: immersive reading of the whole excerpt (30 comparisons) and close reading of 386 aligned HT-MT chunk pairs (772 comparisons), with two readers per book and in alternating order of presentation. Overall, readers find MT "fine", but prefer HT (slightly at excerpt-level 19/30, more clearly at chunk-level 522/772) for its ease, clarity, and immersive nature. Readers' highlights show that MT's quality varies more within one book than HT's does. Crucially, readers cannot reliably tell the two apart (17/30 guess correctly) and tend to prefer the version they believe to be human. Automatic metrics, including LLM-as-a-judge approaches, fail to recover reader preferences and favor MT. We release LAIT (Literary AI Translation), a reader-centered evaluation dataset with 1K reader comments, 2K judgments and preference ratings, and 7.2K span-level annotations, along with our evaluation protocol and supporting interface.
作者Gianmario Voria, Alfonso Cannavale, Andrea De Lucia, Yutaro Kashiwa, Gemma Catolino, Fabio Palomba
Context. Autonomous coding agents are increasingly used in software development, shifting parts of the engineering process to AI assistance. While this automation brings clear benefits, it introduces challenges in governance, traceability, and control over agent behavior. Agent Context Files (ACFs) have emerged as a practical mechanism to guide agents through structured instructions, yet little is known about how these artifacts are maintained and how their evolution relates to code development. Objective. This paper plans to investigate the evolution of ACFs and their role in agent-driven development. Specifically, we (1) classify ACF changes through a taxonomy grounded in software maintenance theory, (2) analyze how different types of changes are associated with code quality outcomes, and (3) examine their temporal patterns across the development lifecycle. Method. We conduct a large-scale mining study combining repositories with ACFs and agent-generated commits. We reconstruct ACF evolution at the commit level, classify changes using a qualitative approach, and analyze their association with code quality metrics. Statistical analyses and hypotheses are used to evaluate differences across maintenance categories, to inform future design of ACFs for governing autonomous coding agents.
展开完整摘要收起摘要↓
Context. Autonomous coding agents are increasingly used in software development, shifting parts of the engineering process to AI assistance. While this automation brings clear benefits, it introduces challenges in governance, traceability, and control over agent behavior. Agent Context Files (ACFs) have emerged as a practical mechanism to guide agents through structured instructions, yet little is known about how these artifacts are maintained and how their evolution relates to code development. Objective. This paper plans to investigate the evolution of ACFs and their role in agent-driven development. Specifically, we (1) classify ACF changes through a taxonomy grounded in software maintenance theory, (2) analyze how different types of changes are associated with code quality outcomes, and (3) examine their temporal patterns across the development lifecycle. Method. We conduct a large-scale mining study combining repositories with ACFs and agent-generated commits. We reconstruct ACF evolution at the commit level, classify changes using a qualitative approach, and analyze their association with code quality metrics. Statistical analyses and hypotheses are used to evaluate differences across maintenance categories, to inform future design of ACFs for governing autonomous coding agents.
The Internet is transitioning from Web3 toward Web4, where autonomous agents serve as independent economic actors. These agents can now hold crypto wallets, execute on-chain trades, and pay for external API calls. This transition calls for a new infrastructure stack capable of supporting key agent operations, including agent-to-tool interaction, agent-to-agent payments, and verifiable agent identity, represented by emerging protocols such as the Model Context Protocol, x402, and EIP-8004. Despite growing industrial interest in these protocols, the real-world Web4 agent ecosystem remains largely underexplored. To bridge this gap, we conduct the first large-scale empirical study of the Web4 ecosystem. Specifically, our study targets three interconnected questions: how Web4 agents are deployed and used in practice; what engineering challenges developers face when building Web4 agents; how current project communities respond to these challenges. To answer these questions, we analyze 99,448 multi-chain identity registrations, 317,596,323 transaction logs, the source code of 341 MCP projects, and 349 filtered GitHub issues. Our findings reveal that autonomous agents have established a highly active machine-to-machine payment economy, processing millions of daily transactions. However, this growth is built on immature infrastructure, including identity/authorization practice, cross-environment operation, and payment interoperability. Our follow-up analysis shows that community responses are visible but unevenly distributed across repositories, and payment interoperability remains the most persistent unresolved bottleneck. Overall, this study reveals a critical gap between the rapid growth of the Web4 agent economy and its fragile underlying infrastructure, highlighting future directions for building a more secure Web4 agent ecosystem.
展开完整摘要收起摘要↓
The Internet is transitioning from Web3 toward Web4, where autonomous agents serve as independent economic actors. These agents can now hold crypto wallets, execute on-chain trades, and pay for external API calls. This transition calls for a new infrastructure stack capable of supporting key agent operations, including agent-to-tool interaction, agent-to-agent payments, and verifiable agent identity, represented by emerging protocols such as the Model Context Protocol, x402, and EIP-8004. Despite growing industrial interest in these protocols, the real-world Web4 agent ecosystem remains largely underexplored. To bridge this gap, we conduct the first large-scale empirical study of the Web4 ecosystem. Specifically, our study targets three interconnected questions: how Web4 agents are deployed and used in practice; what engineering challenges developers face when building Web4 agents; how current project communities respond to these challenges. To answer these questions, we analyze 99,448 multi-chain identity registrations, 317,596,323 transaction logs, the source code of 341 MCP projects, and 349 filtered GitHub issues. Our findings reveal that autonomous agents have established a highly active machine-to-machine payment economy, processing millions of daily transactions. However, this growth is built on immature infrastructure, including identity/authorization practice, cross-environment operation, and payment interoperability. Our follow-up analysis shows that community responses are visible but unevenly distributed across repositories, and payment interoperability remains the most persistent unresolved bottleneck. Overall, this study reveals a critical gap between the rapid growth of the Web4 agent economy and its fragile underlying infrastructure, highlighting future directions for building a more secure Web4 agent ecosystem.
作者Babak Rahmani, Sebastian Dziadzio, Joschka Strüber, Sergio Hernández-Gutiérrez, Matthias Bethge
For most of scientific history, researchers studying behavior could only infer hidden mechanisms from outward actions: an inverse problem that becomes more tractable when observation is augmented by targeted intervention. We pose a computational analogue: given only behavioral traces of an agent in a game environment, can a learner reconstruct the underlying decision program as executable code, and how much does this reconstruction improve with the ability to design controlled experiments? We introduce RevengeBench, a benchmark of 75 LLM generated, Elo-calibrated policies across five game environments, drawn from CodeClash tournament trajectories. The learner observes the hidden target policy play against sampled opponents and designs behavioral probes in the form of custom opponent policies that elicit informative behavior. It then submits an executable hypothesis, which is evaluated using continuous action-distance metrics. We further validate that recovered code carries informative signal in downstream player-versus-player tournaments. Across twelve frontier LLMs, recovery quality varies substantially (34 to 72% of initial distance closed), with reconstructed policies yielding measurable competitive advantage, particularly for weaker models that otherwise struggle to design effective counter-strategies. Our benchmark positions behavioral recovery of programmatic policies as a tractable inverse problem in code-space, opening a path to opponent modeling, policy interpretability, and the broader question of inferring latent mechanisms from observations.
展开完整摘要收起摘要↓
For most of scientific history, researchers studying behavior could only infer hidden mechanisms from outward actions: an inverse problem that becomes more tractable when observation is augmented by targeted intervention. We pose a computational analogue: given only behavioral traces of an agent in a game environment, can a learner reconstruct the underlying decision program as executable code, and how much does this reconstruction improve with the ability to design controlled experiments? We introduce RevengeBench, a benchmark of 75 LLM generated, Elo-calibrated policies across five game environments, drawn from CodeClash tournament trajectories. The learner observes the hidden target policy play against sampled opponents and designs behavioral probes in the form of custom opponent policies that elicit informative behavior. It then submits an executable hypothesis, which is evaluated using continuous action-distance metrics. We further validate that recovered code carries informative signal in downstream player-versus-player tournaments. Across twelve frontier LLMs, recovery quality varies substantially (34 to 72% of initial distance closed), with reconstructed policies yielding measurable competitive advantage, particularly for weaker models that otherwise struggle to design effective counter-strategies. Our benchmark positions behavioral recovery of programmatic policies as a tractable inverse problem in code-space, opening a path to opponent modeling, policy interpretability, and the broader question of inferring latent mechanisms from observations.
作者Yuxin Wang, Paul Thomas, Zhiwei Yu, Yuan Gao, Saeed Hassanpour, Soroush Vosoughi, Robert Sim, Nick Craswell
Prior research on memory mechanism in RAG-based conversational system has emphasized how memory is stored and retrieved. However, far less is known about how memories with different functional roles influence response quality. Specifically, how they shape an agent's responses under varying conversational contexts and whether they lead to substantively different response behaviors. Existing evaluations in conversational system are also largely reference-based, insufficiently capturing the nuances in responses that may address users' preferences differently. In this work, we probe the impact of different memory types in shaping agents' responses. We present a fine-grained taxonomy of conversational memory, classify retrieved memories into different role types, and design a user-centric evaluation framework that simulates user perspectives. Through comparative experiments on long-term datasets and frontier LLMs, our analysis reveal many differentiated effects of memories: e.g., clarifying memory improves responses' factual accuracy and constraint awareness, making them more correct and personalized; irrelevant memory reduces topic relevance and degrades constraint awareness. Despite the power of frontier LLMs, these findings shed light on how different memory types can be leveraged to produce more personalized responses and inspire further research in this direction.
展开完整摘要收起摘要↓
Prior research on memory mechanism in RAG-based conversational system has emphasized how memory is stored and retrieved. However, far less is known about how memories with different functional roles influence response quality. Specifically, how they shape an agent's responses under varying conversational contexts and whether they lead to substantively different response behaviors. Existing evaluations in conversational system are also largely reference-based, insufficiently capturing the nuances in responses that may address users' preferences differently. In this work, we probe the impact of different memory types in shaping agents' responses. We present a fine-grained taxonomy of conversational memory, classify retrieved memories into different role types, and design a user-centric evaluation framework that simulates user perspectives. Through comparative experiments on long-term datasets and frontier LLMs, our analysis reveal many differentiated effects of memories: e.g., clarifying memory improves responses' factual accuracy and constraint awareness, making them more correct and personalized; irrelevant memory reduces topic relevance and degrades constraint awareness. Despite the power of frontier LLMs, these findings shed light on how different memory types can be leveraged to produce more personalized responses and inspire further research in this direction.
Large Language Models (LLMs) are increasingly used for automated vulnerability detection, but it remains unclear how program structure and semantics should be represented for LLM-based reasoning. Most prompting-based approaches provide raw source code, implicitly assuming that more source-level context gives the model better evidence. This paper challenges that assumption through RepBench, an empirical benchmark comparing raw source code with static-analysis-based program representations. RepBench converts real-world C/C++ vulnerability testcases into multiple representations: raw source, Abstract Syntax Trees (ASTs), Control-Flow Graphs (CFGs), Program Dependence Graphs (PDGs), their combinations, and an auxiliary track of enriched PDGs (ePDGs). Using a curated PrimeVul-derived corpus of 107 Joern-based testcases across five CWE categories, we evaluate ten representation variants under a fixed Chain-of-Thought and structured-output protocol, plus 19 additional ePDG cases generated through VulChecker/Hector. Representation choice substantially affects LLM vulnerability reasoning. The strongest variant, AST+PDG, achieves 83.2% accuracy, compared with 53.5% for raw source. At the family level, graph-only prompts outperform both source-only and source-plus-graph prompts while requiring far less prompt overhead. These results reveal a context dilution effect: adding raw source code to compact structural graph evidence can degrade reasoning by making vulnerability-relevant evidence less salient. Overall, our findings show that carefully selected structural representations offer a better accuracy-overhead tradeoff than simply giving LLMs more raw input, and suggest that static analysis can serve as an effective prompt-construction layer for security-focused LLM reasoning.
展开完整摘要收起摘要↓
Large Language Models (LLMs) are increasingly used for automated vulnerability detection, but it remains unclear how program structure and semantics should be represented for LLM-based reasoning. Most prompting-based approaches provide raw source code, implicitly assuming that more source-level context gives the model better evidence. This paper challenges that assumption through RepBench, an empirical benchmark comparing raw source code with static-analysis-based program representations. RepBench converts real-world C/C++ vulnerability testcases into multiple representations: raw source, Abstract Syntax Trees (ASTs), Control-Flow Graphs (CFGs), Program Dependence Graphs (PDGs), their combinations, and an auxiliary track of enriched PDGs (ePDGs). Using a curated PrimeVul-derived corpus of 107 Joern-based testcases across five CWE categories, we evaluate ten representation variants under a fixed Chain-of-Thought and structured-output protocol, plus 19 additional ePDG cases generated through VulChecker/Hector. Representation choice substantially affects LLM vulnerability reasoning. The strongest variant, AST+PDG, achieves 83.2% accuracy, compared with 53.5% for raw source. At the family level, graph-only prompts outperform both source-only and source-plus-graph prompts while requiring far less prompt overhead. These results reveal a context dilution effect: adding raw source code to compact structural graph evidence can degrade reasoning by making vulnerability-relevant evidence less salient. Overall, our findings show that carefully selected structural representations offer a better accuracy-overhead tradeoff than simply giving LLMs more raw input, and suggest that static analysis can serve as an effective prompt-construction layer for security-focused LLM reasoning.
Recent LLM role-playing systems build character agents from novels by extracting characters, scenes, and relations. Yet long-narrative role-playing suffers from two failures: Factual Overreach, where shared retrieval or parametric memory lets a character use facts outside its perspective, and Stylistic Monotony, where profile descriptions flatten a character into a fixed voice. To address these failures, we propose REVERIEMEM, a three-layer memory architecture for book-based character agents. The episodic layer stores first-person scene memories; the semantic layer stores visibility-tagged facts; and the personality layer stores situation-dependent speech and behaviour patterns. For evaluation, we construct KBF-QA, a 4,386-question benchmark over eight novels for testing knowledge boundaries. REVERIEMEM improves Knowledge Boundary Fidelity by 34.6 percentage points over the strongest prior method. On BOOKWORLD's five-dimension pairwise narrative protocol, REVERIEMEM achieves a ~ 79% win rate, suggesting that perspective-bounded memory improves both boundary fidelity and character-grounded narrative generation.
展开完整摘要收起摘要↓
Recent LLM role-playing systems build character agents from novels by extracting characters, scenes, and relations. Yet long-narrative role-playing suffers from two failures: Factual Overreach, where shared retrieval or parametric memory lets a character use facts outside its perspective, and Stylistic Monotony, where profile descriptions flatten a character into a fixed voice. To address these failures, we propose REVERIEMEM, a three-layer memory architecture for book-based character agents. The episodic layer stores first-person scene memories; the semantic layer stores visibility-tagged facts; and the personality layer stores situation-dependent speech and behaviour patterns. For evaluation, we construct KBF-QA, a 4,386-question benchmark over eight novels for testing knowledge boundaries. REVERIEMEM improves Knowledge Boundary Fidelity by 34.6 percentage points over the strongest prior method. On BOOKWORLD's five-dimension pairwise narrative protocol, REVERIEMEM achieves a ~ 79% win rate, suggesting that perspective-bounded memory improves both boundary fidelity and character-grounded narrative generation.
作者Yonathan Michael, Mohamad Alansari, Natnael Takele, Andreas Henschel, Naoufel Werghi
Conventional vision-language models are largely object-centric, focusing on detecting and describing individual entities. In safety-critical X-ray baggage screening, however, threat often emerges not from a single object but from the functional compatibility of spatially dispersed components, such as batteries, detonators, and explosive charges. We formalize this setting as compositional threat reasoning, where risk is modeled as a relational property of grounded regions rather than an independent detection outcome. We introduce Falcon, a multimodal framework that abstracts segmentation-aware region features into a structured safety state capturing component presence, pairwise functional compatibility, and scene-level risk. This structured representation is injected into the language model as an explicit intermediate interface, encouraging relationally consistent and safety-aware reasoning. To evaluate this problem, we present Falcon-X, a benchmark that unifies dense grounding with structured supervision over component completeness and risk inference in cluttered X-ray imagery. Experiments show that while existing multimodal models adapt to appearance, they struggle with compositional safety reasoning. Falcon improves functional grounding and produces more coherent threat assessments, establishing compositional safety reasoning as a distinct evaluation paradigm for multimodal systems.
展开完整摘要收起摘要↓
Conventional vision-language models are largely object-centric, focusing on detecting and describing individual entities. In safety-critical X-ray baggage screening, however, threat often emerges not from a single object but from the functional compatibility of spatially dispersed components, such as batteries, detonators, and explosive charges. We formalize this setting as compositional threat reasoning, where risk is modeled as a relational property of grounded regions rather than an independent detection outcome. We introduce Falcon, a multimodal framework that abstracts segmentation-aware region features into a structured safety state capturing component presence, pairwise functional compatibility, and scene-level risk. This structured representation is injected into the language model as an explicit intermediate interface, encouraging relationally consistent and safety-aware reasoning. To evaluate this problem, we present Falcon-X, a benchmark that unifies dense grounding with structured supervision over component completeness and risk inference in cluttered X-ray imagery. Experiments show that while existing multimodal models adapt to appearance, they struggle with compositional safety reasoning. Falcon improves functional grounding and produces more coherent threat assessments, establishing compositional safety reasoning as a distinct evaluation paradigm for multimodal systems.