每日文章

Intrinsic self-correction (SC) aims to improve large language model outputs by prompting a model to revisit its own initial answer without external feedback. Recent studies have questioned the reliability of this approach, showing that models often struggle to judge whether their initial responses are correct. In this work, we take a task-sensitive view of SC. Rather than asking whether it works in general, we examine settings where SC may operate through different mechanisms: verifying explicit constraints, revisiting a complex reasoning process, or providing a second opinion over competing strategies in word-game tasks. Across multiple benchmarks and models, we find that SC can yield consistent performance gains when the underlying task structure facilitates these modes of revision. These results suggest that SC is best understood as a task-dependent inference-time strategy whose usefulness depends on the role the revision stage can play in a given task, rather than as a uniformly reliable method for improving initial model outputs.

Operating in complex real-world environments requires robots to understand their surroundings on a functional semantic level. This demands a detailed multi-layer world model capturing the complex relations of its surroundings. Hierarchical 3D scene graphs address this challenge by integrating geometric, semantic, and relational data within a unified spatial framework. However, current 3D scene graph approaches often restrict themselves to rigid structures of pre-determined relationship classes, mostly neglecting important semantic connections, like causal connections or environmental contexts. This paper explores the potential of foundation models to build forests of 3D scene graphs with open semantic relationships to improve scene understanding and robotic task execution. We propose a method where instance-specific concept-nodes and relationships are first identified by a VLM and extended upon by a LLM, inferring broader, more abstract concept-nodes and relationships through reasoning. These object-nodes, concept-nodes, and relationships are then assembled into a forest of hierarchical 3D scene graphs, enhanced with concept-nodes to represent abstract concepts. Evaluations were conducted on the uHumans2 and ScanNet indoor dataset, validating the accuracy and relevance of the generated relationships. Downstream suitability of scene-graph forests for robotics applications is demonstrated in an open-vocabulary object-retrieval task utilizing both ScanNet data and a real-world indoor deployment using a Boston Dynamics Spot. This paper leverages foundation models to create more expressive, semantically deep 3D hierarchical scene graphs and demonstrates their potential to advance semantic and environmental understanding in robotics.

Recent large language models (LLMs) are good at general text generation, but it is still hard to use them for domain-specific data generation because the output must follow strict formatting and structural rules. Unlike open-ended tasks such as question answering or translation, domain-specific generation must be both semantically correct and compliant with existing guidelines and standards. In this work, we study the nationwide interoperability problem of utility power outage reports in the United States. In practice, outage reports need to be machine-readable (e.g., JSON or XML) and must strictly follow requirements from energy-sector regulatory bodies. To address this problem, we propose POTracker, an optimized LLM for power outage report generation. We fine-tune Qwen2.5-7B-Instruct using our proposed objective. The key contribution is a new loss function, POTrackerLoss, that considers both textual similarity and structural (tag) similarity between the generated report and the ground-truth report. We evaluate POTracker on a dataset of 1,000 power outage reports and compare it with five well-known fine-tuning methods and one rule-based XML conversion method. Results show that POTracker outperforms other fine-tuning approaches, improving overall accuracy by up to 51% and reaching 86.47% structural accuracy for generated power outage reports. In addition, we conduct a human study to assess the quality of the ground-truth standard reports, where domain experts assign the generated labels an average score of 4.03 on a 0--5 scale.

Language model reasoning can be substantially improved at test time via scaffolds that scale inference compute across different primitives -- sequential reasoning within a trace, independently sampled parallel traces, and aggregation of multiple reasoning traces into a final response. During post-training, however, language models are optimized only for sequential reasoning within a single trace. We introduce Sequential-Parallel-Aggregative Reinforcement Learning (SPIRAL), a framework in which a language model is trained to use all three primitives, as part of a unified inference compute pipeline. Concretely, the language model first samples a set of independent traces in parallel, each produced through sequential chain-of-thought reasoning, and then generates a final aggregation trace conditioned on those traces; all components are optimized end-to-end against the reward of the final aggregated response. To train this system, SPIRAL uses set reinforcement learning to teach models to produce a set of traces that are collectively useful for an aggregator and standard reinforcement learning to teach models to aggregate the set into improved final responses. Our experiments on reasoning tasks show that SPIRAL effectively scales with inference compute, outperforming GRPO by up to 11\(\times\) scaling efficiency and 15% higher performance when all three compute primitives are scaled.

This paper presents our algorithmic innovations for the NVIDIA Nemotron Model Reasoning Challenge, focusing on Bit Manipulation Puzzles. In this task, the objective is to discover a hidden logical rule transforming input binary strings to outputs, then apply it to unseen inputs. Large Language Models (LLMs) notoriously struggle here; traditional methods force them to simulate complex boolean logic and arithmetic, leading to hallucinations. Furthermore, the search space of bitwise operations (combinations of shifts, rotations, and logic gates) suffers from a severe combinatorial explosion. To overcome this computational intractability, we present a novel approach that abandons arithmetic logic entirely in favor of string similarity, structured search, and autonomous error recovery. Our core contributions are: 1. Bases and Truth Table Formulation: We reframe logic-gate deduction into a base-selection task, leveraging string similarity (minimal bit flips) to isolate primitive transformations ("bases") and deduce truth tables without complex arithmetic. 2. Backtracking DFS and Error Recovery: We formalize a search process that tests candidate bases, detects logical collisions across examples, and backtracks upon failure to perform robust error recovery. 3. Bit Tokenization and Interactive Reasoning SFT: We force the tokenizer to encode binary strings as individual single-bit tokens. We use dynamic masking to simulate external oracle feedback, training the model to hypothesize, self-evaluate, and backtrack natively. Evaluated on bit manipulation puzzles, our approach achieved over 96% validation accuracy. This represents the highest performance in this category, driving our 7th Place overall finish in the contest.

Deploying foundation models for robot control raises a central challenge: the expressive power that enables rich, multimodal perception also makes these models opaque and difficult to analyze formally, rendering them intractable for existing verification tools. In this paper, we present FEARL (Foundation-Enabled Assured Robot Learning), a framework that addresses this tension through a modular architectural decomposition. FEARL separates the policy into a large Controller (C) responsible for high-dimensional perception and task reasoning, and a small Safety module (S) that receives low-dimensional observations from dedicated safety sensors together with a bounded context embedding from C and produces the final action. Since many robot safety requirements, such as collision avoidance and workspace boundary constraints, can be expressed over these safety sensor observations, formal verification can be applied to S rather than to the full foundation-model backbone. This makes formal analysis tractable with existing tools while preserving the Controller's expressive power for task reasoning. To show that the decomposed policy remains capable of solving diverse tasks, we evaluate FEARL on three simulated robotic domains using multiple Controller backbones and training procedures, including pretrained off-the-shelf vision-language-action models. We further transfer the learned policy from one of our simulated tasks to a physical robot, suggesting that the low-dimensional safety interface supports practical sim-to-real transfer.

Clinical value sets define the standardized terminology codes used in quality measurement, phenotyping, cohort construction, and clinical decision support. The recently introduced Retrieval-Augmented Set Completion (RASC) benchmark showed that direct zero-shot large language model (LLM) generation is poorly suited to this task: clinical code systems are large, version-controlled, and not reliably memorized by language models. We study a stage-wise alternative in which candidate-pool construction is optimized for recall and a constrained LLM adjudicator is optimized for candidate selection. On the full 3,744-value-set RASC test split, Qwen3-based retrieval with vocabulary-aware expansion and code-display rescue retrieval increases candidate-pool recall from the original RASC retrieval baseline of 0.553 to 0.730; on the held-out-publisher stratum, pool recall is 0.655. The higher-recall pool alone is not sufficient: applying the original SAPBert cross-encoder to this expanded pool gives full-test macro F1 of 0.287 and held-out-publisher macro F1 of 0.233. Replacing the stage-2 selector with blinded GPT-5 adjudication over the same pool increases full-test macro F1 to 0.549 and held-out-publisher macro F1 to 0.533. These results show that retrieval-constrained LLM adjudication can substantially improve value set completion while preserving the safety constraint that all returned codes must come from an auditable candidate pool.

Memory systems are essential for personalized Large Language Models (LLMs). However, existing retrieval methods in these systems primarily rely on semantic similarity, potentially missing logically critical memories with limited semantic overlap. Current benchmarks remain inadequate for evaluating this problem. To address this gap, we construct IMLogic, the first high-quality benchmark targeting implicit logical memory retrieval in long-dialogue scenarios. Motivated by this challenge, we introduce root memory, a structured, decision-preserving representation that distills reusable personalized logic from long-term user histories. We then propose RootMem, a plug-and-play framework that first distills raw histories into structured root memories and then uses an LLM-based router to activate logically relevant ones, complementing semantic retrieval with personalized decision logic. Extensive experiments demonstrate that RootMem significantly outperforms the strongest retrieval baselines and consistently boosts the accuracy of existing memory agents. Our benchmark and codes will be available at https://anonymous.4open.science/r/IMLogic-DBB3.

Enterprise agents increasingly operate inside workspaces: they read heterogeneous files, invoke tools, and deliver business artifacts. We introduce EnterpriseClawBench, an enterprise agent benchmark constructed from proprietary, real-world agent sessions. Starting from a large archive of workplace sessions, the EnterpriseClawBench produces 852 reproducible tasks, each paired with recovered fixtures, rewritten prompts, role classes, skill subclasses, hard rules, and semantic rubrics. Because the sessions contain internal enterprise content, we do not release the benchmark data; instead, our reusable contribution is the construction and evaluation protocol. On EnterpriseClawBench, the best configuration reaches only 0.663 (Codex with GPT-5.5). These results show that enterprise agent evaluation must report harness--model combinations, artifact delivery, visual quality, cost, runtime, and skill-transfer behavior, rather than collapsing performance into a single score. Code: https://github.com/FrontisAI/EnterpriseClawBench

AdamW is the de facto optimizer for training large language models (LLMs), yet the theory behind it still lives mostly in finite-variance regimes. This is increasingly unsatisfying, as empirical evidence indicates that stochastic gradient noise in LLM pretraining is typically heavy-tailed. Recent work shows that sign-based optimizers such as Lion and Muon achieve sharp heavy-tailed rates, and that AdaGrad can also converge under heavy-tailed noise. However, no rigorous convergence theory for AdamW has yet been established in this regime. Can AdamW converge under the same heavy-tailed assumptions, or does its second-moment accumulator create a genuine obstruction? We formulate this as an open problem, prove a positive weighted-metric benchmark, and give a corridor lower-bound mechanism showing how denominator memory can hide large gradients.

Driving VLA models incorporating Chain-of-Thought (CoT) reasoning are attractive because they leverage pretrained VLM representations and expose intermediate decisions in natural language, yet current rationales often lack the step-by-step decision semantics needed to keep the rationale causally connected to the planned motion. We introduce Neuro-Symbolic Drive, a neuro-symbolic driving framework that supervises a driving VLA with rule-grounded reasoning traces extracted directly from classical rule-based planners. Our key observation is that rule-based planners are symbolic AI systems that already function as executable reasoning engines: they reason about active safety constraints, search over candidate maneuvers, and select a final trajectory. We instrument these planners in simulation to capture both the executed trajectory and the internal decision trace at each rule-evaluation step. Each trace is serialized into structured rule-grounded reasoning and paired with the trajectory to fine-tune Qwen3.5-4B as a driving VLA. Because these traces are derived directly from the planner states that determine the action, they ensure reasoning is structurally coupled to motion generation by construction, rather than by post-hoc alignment. On our simulator-generated benchmark, detailed rule-grounded reasoning reduces ADE@3s from 0.47 to 0.26 and miss rate from 8.30% to 6.40% under three-camera perception, and from 0.54 to 0.26 and 10.13% to 5.99% under eight-camera perception. Neuro-Symbolic Drive thus converts neuro-symbolic planning logic into structured supervision. Code base: https://github.com/XiangboGaoBarry/Neural-Symbolic-Drive.

AI agents increasingly act through external tools: they read private data, construct structured payloads, submit write requests, export records, and coordinate workflows across application boundaries. Existing authorization mechanisms usually ask whether an integration credential, app, or token can call a tool. That question is necessary but incomplete. A tool call can be authorized by static credentials and still be unjustified by the user's current request. For example, a credential that can read and export records should not expose export authority when the user only asked for a bounded summary, and a model-generated delete call should not execute merely because the integration has a delete scope. This paper proposes Intent-Governed Access Control (IGAC), a server-side authorization layer that treats the user's expressed intent as a monotone, auditable policy attribute for AI-agent tool use. IGAC introduces intent certificates, session-scoped policy narrowing, intent-aware manifest filtering, and intent-tool-payload consistency checks. The central invariant is that user intent may only reduce the authority granted by static integration policy; it never expands scopes, data policy, tenant boundaries, or review requirements. We map IGAC onto OpenPort, an existing governance substrate that already implements authorization-dependent discovery, scope and ABAC-style policy checks, draft-first writes, preflight impact binding, state-witness checks, idempotency, stable reason codes, and audit.

Verilog debugging remains one of the most time-consuming stages in digital circuit design. Recent advances in Large Language Models (LLMs) have enabled automated debugging; however, most existing approaches rely solely on test outputs and compiler feedback in an end-to-end manner, limiting their effectiveness on complex bugs. A key challenge is that the root cause of an error may be far removed from its observable outputs, making it difficult for LLMs to trace long dependency chains in code. This challenge is further exacerbated in large codebases, where long context lengths hinder efficient reasoning. To address these limitations, we propose VeriPilot, an LLM-powered debugging framework that leverages golden reference models to enable fine-grained bug localization and repair. VeriPilot goes beyond output-level comparison by aligning internal variable semantics between the Verilog design and its corresponding golden model through LLM-based analysis. It then performs step-by-step signal tracing using Control-Data-Flow Graphs (CDFGs) derived from static analysis, identifying a minimal set of suspicious code regions along with their correct counterparts from the golden model. These structured insights are subsequently provided to the LLM to guide reasoning and automated code repair. Experimental results on the Comprehensive Verilog Design Problems (CVDP) benchmark from NVIDIA demonstrate that VeriPilot improves the repair success rate of GPT-4o from 54.3\% to 85.71\%, significantly enhancing both bug localization accuracy and repair effectiveness for complex Verilog designs. The source code and benchmark are publicly available at Github https://github.com/YihanWn/VeriPilot.git.

Knowledge graphs (KGs) are increasingly used as structured context for Large Language Models (LLMs), but industrial KG-RAG systems often need to integrate public and domain-specific KGs constructed from heterogeneous databases. This integration relies on Entity Alignment (EA), where lexical matching alone is insufficient under predicate-name variation and incomplete local neighborhoods. We address EA for KG integration by constructing a pairwise EA dataset and proposing two complementary modules: Predicate Importance Estimation (PIE) and Decoupled Rationale-Score Distillation (DRSD). PIE is a compact embedding-based approach that removes the subject information from each 1-hop triple, encodes the resulting subjectless triples, and aggregates them with learnable predicate-importance weights to build predicate-aware entity embeddings. DRSD trains a distilled small language model (SLM) with pseudo-answers produced by a teacher LLM through distinct prompts. By converting binary EA labels into text-based supervision and decoupling confidence-score estimation from label-consistent rationales, DRSD enables the SLM to learn task-specific reasoning while retaining a less label-biased confidence signal. Experiments show that PIE and DRSD improve EA classification. Moreover, because DRSD decouples confidence-score estimation from the decision, a discrepancy between the two flags an uncertain prediction for human review, thereby enabling a practical discrepancy between automatic acceptance and human-in-the-loop verification.

Clinical audio diagnosis in low-resource settings requires models that identify conditions from minimal examples without large annotated corpora. We propose Federated Self-Contextualization (FSC), a multimodal language model framework for in-context clinical audio diagnosis across federated hospital clients. FSC constructs pseudo-label episodes via unsupervised clustering of audio representations, bypassing scarce real diagnostic labels, and enables contextual reasoning from support-query pairs. Our progressive three-stage pipeline first aligns audio embeddings with the language model via caption-based pretraining, then adapts it for episodic in-context inference through federated optimization. At test time, given a small labeled support set, the model diagnoses an unseen query through multimodal reasoning. On held-out respiratory and cardiac conditions, FSC achieves 71.6% accuracy in 2-way 2-shot evaluation, outperforming audio-language baselines by over 9%.

Closed-loop Auto Research extends automated machine learning from fixed-dataset fitting to changing the research workflow, with language-model agents editing representations and model code and acquiring external evidence. Molecular property prediction spans many small endpoints. We ask whether this action space yields improvements generalizing beyond the validation signal selecting them. We isolate three Auto Research axes, features, models, and external evidence, under a file-level ablation lock attributing each gain to one axis over a strong baseline. Across 36 endpoints in three benchmark suites we score each selected configuration once on a held-out test whose labels the search never read. A routed pipeline taking each endpoint's best validation axis reaches positive held-out gains of 0.013, 0.011, and 0.042, the transferable axis differing by suite, data on TDC, model on Polaris, feature and model on MoleculeNet. The largest model-search gain falls from 0.041 on validation to 0.003 on test, while curated data reaches 0.022 but negative 0.019 on test, two non-transfer signatures. Curated external data raises held-out CYP2C9-substrate performance by 0.17 and half-life by 0.08, admitted through a contamination filter rejecting same-source files overlapping 64 to 89 percent of test structures, necessary but not sufficient for transfer. A matched-trial automated machine learning control did not reproduce the agent's code-level model intervention, reaching 0.006 against 0.042, and the pipeline stays competitive with an 84M-parameter pretrained 3D model on the shared training split. The experiments stay within molecular property prediction, but separating discovery from held-out certification is a domain-agnostic lesson for any closed-loop system optimising a proxy for a held-out quantity.

Exact-match retrieval recall is often used as a proxy for whether a retriever supplies useful policy context to a downstream decision model. We test this proxy for pre-action policy classification in tau-bench using Qwen2.5-3B/7B classifiers. Under gold-policy conditioning, a compact structured state improves macro-F1 over raw trajectories by 0.13-0.17 after tuning. We then replace the benchmark-designated policy clause with the top-ranked clause retrieved from decision-time context. Although the exact governing clause is retrieved at rank 1 for only 7% of airline states, the primary 3B classifier obtains macro-F1 0.58 with retrieved clauses versus 0.60 with gold clauses (Delta=-0.02, task-cluster 95% CI [-0.23,+0.21]); mismatched-policy and no-policy controls score 0.32 and 0.21. We do not detect a macro-F1 difference between retrieved and gold clauses in this configuration, although the interval remains too wide to establish non-inferiority. The same qualitative pattern appears with a second retriever and at 7B, while varying across fine-tuning configurations. These results indicate that exact-match clause recall can underestimate downstream policy utility in this benchmark setting, motivating evaluation with retrieved policies in the classification loop rather than recall alone.

Statistical analysis is a broad, complex field requiring both domain knowledge and tool proficiency. While prior work has evaluated large language models (LLMs) in this domain, existing benchmarks remain limited in scope and format. To bridge this gap, we introduce StatABench (Statistical AnalysisBenchmark), a benchmark designed to systematically assess LLMs' statistical analysis capabilities. StatABench comprises two complementary components: Stat-Closed, containing 404 questions across 18 statistical topics in multiple formats (multiple-choice, fill-in-the-blank, decision-making, and practical application), and Stat-Open, featuring 30 complex open-ended modeling tasks adapted from professional competitions. We evaluate diverse LLMs using the LangChain MCP framework and multiple data science agents, and assess Stat-Open solutions via a validated LLM-as-Judge protocol. Experiments show that even GPT-5.1 achieves only 68.6% on Stat-Closed, while the best open-source model reaches 60.6%. On Stat-Open, the top agent framework scores 61.86 on average. These results reveal the gap between current LLMs and reliable statistical analysis, highlighting persistent challenges in tool-grounded reasoning, methodological decision-making, and end-to-end statistical modeling.

Computer-use agents (CUAs) now act on a user's behalf across personal applications such as email, calendars, and to-do lists. This cross-application access is useful, but it also creates a privacy risk that has been largely overlooked: when an agent works in one context, it can pull in information from another that is inappropriate in that context. Hence, we introduce AgentCIBench, an evaluation harness that turns this risk into executable, deterministically scored scenarios. We target three common failure modes in CUAs: visual co-location, where the agent pulls in prohibited items that sit next to the task target in the UI; task-ambiguity overshare, where the agent dumps dense personal state in response to an under-specified prompt; and recipient misalignment, where the agent sends content to an addressee for whom it is inappropriate. We evaluate 15 frontier agents and find a surprisingly high failure rate: 11 of 15 leak on more than 50% of scenarios, with an average leakage of 67.9%, and the same failures persist when agents act end-to-end in the environment to complete the task. We release AgentCIBench to encourage the development of safer computer-use agents and position contextual disclosure testing as a pre-deployment safety check.

Finance Agent v2 (by Vals AI) has emerged as the reference benchmark for evaluating both Anthropic Claude and OpenAI ChatGPT frontier language models on financial tasks. However, it narrowly deals with periodic reporting from publicly traded companies (SEC 10-K and 10-Q filings), and its agentic harness relies on naive, unenriched chunk retrieval. Neither the task design nor the retrieval approach addresses the distinct challenges of IPO due diligence. SEC S-1 filings combine historical financial statements, governance structures, pro forma and common-control accounting treatments, capital-formation narratives, and underwriting-sensitive risk disclosures within substantially longer documents than typical periodic filings. That is why we introduce IPO Finance Agent, which extends the Finance Agent v2 framework along two directions: task domain and retrieval architecture. During our experiments, the original Finance Agent v2 harness basically failed to deliver any output related to the SpaceX S-1 filing, due to document length. We therefore had to improve the agentic harness with contextual retrieval, a more realistic and industry-standard approach for long documents. We also built a dataset of 1,000 IPO-diligence questions, and publicly release 70 questions on the SpaceX (SPCX) S-1 filing to support reproducibility, while the remainder are held private to guard against benchmark contamination. In addition, we introduce an evaluator-optimizer pipeline to automatically generate evaluation rubrics for the benchmark: candidate facts are extracted from model answers, consolidated into draft criteria, then automatically audited for omissions, hallucinations, mistiered items, and redundancy, with LLM feedback driving iterative repair, targeted enrichment, and deduplication. Human experts only review final rubrics before deployment. Results show that the best-performing evaluated model, Alibaba Qwen 3.7 Max, reaches 79.4% accuracy at 0.30 USD per query, and the most cost-efficient model on the resulting Pareto frontier, Xiaomi MiMo-2.5 Pro, reaches slightly lower accuracy (76.8%) at 0.05 USD per query. Both exceed the current Finance Agent v2 leaderboard ceiling-Google Gemini 3.5 Flash at 57.9% for 2.51 USD per querywhile undercutting even FABv2's cheapest entry (MiniMax M3: 48.3% at 0.32 USD) on cost-efficiency. Code and data are released on GitHub: https://github.com/benstaf/ipoagent