{ "generated_at": "2026-06-18T14:03:08.733612+08:00", "timezone": "Asia/Shanghai", "lookback_hours": 24, "sorting": { "default_sort_by": "hybrid", "summary": "hybrid (relevance first, published_at tie-break)", "weights": { "title_match_weight": 40, "summary_match_weight": 18, "doi_weight": 12, "pdf_weight": 8, "rich_summary_weight": 6, "metadata_weight": 4, "multi_source_weight": 10, "freshness_weight_cap": 24 }, "feeds": [ { "name": "LM", "sort_by": "hybrid" }, { "name": "Agent Runtime Security", "sort_by": "hybrid" }, { "name": "Terminal and SWE Agents", "sort_by": "hybrid" } ] }, "highlights": [ "主题「LLM」：命中 16 篇，覆盖 LM、Agent Runtime Security，代表论文包括 《Enhancing Decision-Making with Large Language Models through Multi-Agent Fictitious Play》、《IndicContextEval: A Benchmark for Evaluating Context Utilisation in Audio Large Language Models Across 8 Indic Languages》。", "主题「Language Model」：命中 15 篇，覆盖 LM、Agent Runtime Security，代表论文包括 《Enhancing Decision-Making with Large Language Models through Multi-Agent Fictitious Play》、《IndicContextEval: A Benchmark for Evaluating Context Utilisation in Audio Large Language Models Across 8 Indic Languages》。", "主题「Benchmark」：命中 2 篇，覆盖 LM、Terminal and SWE Agents，代表论文包括 《G-IdiomAlign: A Gloss-Pivoted Benchmark for Cross-Lingual Idiom Alignment》、《Data Intelligence Agents: Interpreting, Modeling, and Querying Enterprise Data via Autonomous Coding Agents》。", "主题「Agent」：命中 1 篇，覆盖 Terminal and SWE Agents，代表论文包括 《Data Intelligence Agents: Interpreting, Modeling, and Querying Enterprise Data via Autonomous Coding Agents》。" ], "focus_items": [], "action_items": [], "topic_sections": [ { "name": "LLM", "paper_count": 16, "feed_names": [ "LM", "Agent Runtime Security" ], "paper_titles": [ "Enhancing Decision-Making with Large Language Models through Multi-Agent Fictitious Play", "IndicContextEval: A Benchmark for Evaluating Context Utilisation in Audio Large Language Models Across 8 Indic Languages", "A Technical Taxonomy of LLM Agent Communication Protocols", "Quantifying and Auditing LLM Evaluation via Positive--Unlabeled Learning", "Trade-offs in Medical LLM Adaptation: An Empirical Study in French QA", "X+Slides: Benchmarking Audience-Conditioned Slide Generation", "CAPRA: Scaling Feedback on Software Architecture Deliverables with a Multi-Agent LLM System", "Beyond Tokenization: Direct Timestep Embedding and Contrastive Alignment for Time-Series Question Answering", "Dango: A Strictly L1-Only Large Language Model for Studying Second Language Acquisition", "G-IdiomAlign: A Gloss-Pivoted Benchmark for Cross-Lingual Idiom Alignment", "Beyond Safe Data: Pretraining-Stage Alignment with Regular Safety Reflection", "AdsMind: A Physics-Grounded Multi-Agent System for Self-Correcting Discovery of Adsorption Configurations on Heterogeneous Catalyst Surfaces", "OpenAnt: LLM-Powered Vulnerability Discovery Through Code Decomposition, Adversarial Verification, and Dynamic Testing", "Seeing Before Reasoning: Decoupling Perception and Reasoning for Shortcut-Resilient Multimodal On-Policy Self-Distillation", "Diffusion-Proof: Recipe for Formal Theorem Proving Beyond Auto-Regressive Generation", "CodeSentinel: A Three-Layer Defense Against Indirect Prompt Injection in Code Contexts" ], "key_points": [ "《Enhancing Decision-Making with Large Language Models through Multi-Agent Fictitious Play》〔评测 / 方法〕：Large language model (LLM)-based multi-agent systems (MAS) have demonstrated great potential in solving tasks with execution complexity, by distributing subtas…", "《IndicContextEval: A Benchmark for Evaluating Context Utilisation in Audio Large Language Models Across 8 Indic Languages》〔评测 / 方法〕：AudioLLMs enable speech recognition conditioned on textual prompts such as domain descriptions or entity lists. However, it remains unclear whether these model…" ] }, { "name": "Language Model", "paper_count": 15, "feed_names": [ "LM", "Agent Runtime Security" ], "paper_titles": [ "Enhancing Decision-Making with Large Language Models through Multi-Agent Fictitious Play", "IndicContextEval: A Benchmark for Evaluating Context Utilisation in Audio Large Language Models Across 8 Indic Languages", "A Technical Taxonomy of LLM Agent Communication Protocols", "Quantifying and Auditing LLM Evaluation via Positive--Unlabeled Learning", "Trade-offs in Medical LLM Adaptation: An Empirical Study in French QA", "X+Slides: Benchmarking Audience-Conditioned Slide Generation", "CAPRA: Scaling Feedback on Software Architecture Deliverables with a Multi-Agent LLM System", "Beyond Tokenization: Direct Timestep Embedding and Contrastive Alignment for Time-Series Question Answering", "Dango: A Strictly L1-Only Large Language Model for Studying Second Language Acquisition", "Beyond Safe Data: Pretraining-Stage Alignment with Regular Safety Reflection", "AdsMind: A Physics-Grounded Multi-Agent System for Self-Correcting Discovery of Adsorption Configurations on Heterogeneous Catalyst Surfaces", "OpenAnt: LLM-Powered Vulnerability Discovery Through Code Decomposition, Adversarial Verification, and Dynamic Testing", "Seeing Before Reasoning: Decoupling Perception and Reasoning for Shortcut-Resilient Multimodal On-Policy Self-Distillation", "Diffusion-Proof: Recipe for Formal Theorem Proving Beyond Auto-Regressive Generation", "CodeSentinel: A Three-Layer Defense Against Indirect Prompt Injection in Code Contexts" ], "key_points": [ "《Enhancing Decision-Making with Large Language Models through Multi-Agent Fictitious Play》〔评测 / 方法〕：Large language model (LLM)-based multi-agent systems (MAS) have demonstrated great potential in solving tasks with execution complexity, by distributing subtas…", "《IndicContextEval: A Benchmark for Evaluating Context Utilisation in Audio Large Language Models Across 8 Indic Languages》〔评测 / 方法〕：AudioLLMs enable speech recognition conditioned on textual prompts such as domain descriptions or entity lists. However, it remains unclear whether these model…" ] }, { "name": "Benchmark", "paper_count": 2, "feed_names": [ "LM", "Terminal and SWE Agents" ], "paper_titles": [ "G-IdiomAlign: A Gloss-Pivoted Benchmark for Cross-Lingual Idiom Alignment", "Data Intelligence Agents: Interpreting, Modeling, and Querying Enterprise Data via Autonomous Coding Agents" ], "key_points": [ "《G-IdiomAlign: A Gloss-Pivoted Benchmark for Cross-Lingual Idiom Alignment》〔评测 / 方法〕：Idioms are difficult to transfer across languages due to their non-compositionality and weak surface-form grounding, making literal mappings unreliable. We pre…", "《Data Intelligence Agents: Interpreting, Modeling, and Querying Enterprise Data via Autonomous Coding Agents》〔评测 / 应用 / 方法〕：Production data integration is bottlenecked by repeated, lossy handoffs between data owners, engineers, and analysts who must collaboratively discover, structu…" ] }, { "name": "Agent", "paper_count": 1, "feed_names": [ "Terminal and SWE Agents" ], "paper_titles": [ "Data Intelligence Agents: Interpreting, Modeling, and Querying Enterprise Data via Autonomous Coding Agents" ], "key_points": [ "《Data Intelligence Agents: Interpreting, Modeling, and Querying Enterprise Data via Autonomous Coding Agents》〔评测 / 应用 / 方法〕：Production data integration is bottlenecked by repeated, lossy handoffs between data owners, engineers, and analysts who must collaboratively discover, structu…" ] } ], "template": "zh_daily_brief", "feeds": [ { "name": "LM", "key_points": [ "《Enhancing Decision-Making with Large Language Models through Multi-Agent Fictitious Play》〔评测 / 方法〕：Large language model (LLM)-based multi-agent systems (MAS) have demonstrated great potential in solving tasks with execution complexity, by distributing subtas…", "《IndicContextEval: A Benchmark for Evaluating Context Utilisation in Audio Large Language Models Across 8 Indic Languages》〔评测 / 方法〕：AudioLLMs enable speech recognition conditioned on textual prompts such as domain descriptions or entity lists. However, it remains unclear whether these model…", "《A Technical Taxonomy of LLM Agent Communication Protocols》〔应用 / 方法〕：As large language models (LLMs) advance and multi-agent systems aim to overcome the limits of standalone agents, robust communication protocols are becoming es…", "《Quantifying and Auditing LLM Evaluation via Positive--Unlabeled Learning》〔评测 / 方法〕：Large Language Models (LLMs) are increasingly used as judges for scalable evaluation, yet such LLM--as--a--Judge systems exhibit systematic biases that are dec…", "《Trade-offs in Medical LLM Adaptation: An Empirical Study in French QA》〔评测 / 应用 / 方法〕：The development of large language models (LLMs) has led to an increased focus on their adaptation to specialized domains and languages, yet the effectiveness o…" ], "sort_by": "hybrid", "papers": [ { "title": "Enhancing Decision-Making with Large Language Models through Multi-Agent Fictitious Play", "summary": "Large language model (LLM)-based multi-agent systems (MAS) have demonstrated great potential in solving tasks with execution complexity, by distributing subtasks across cooperative agents. However, this divide-and-conquer paradigm falls short on decision-making tasks that are also prevalent in the real world. These tasks require simultaneous reasoning from the stances of all involved stakeholders whose decisions are mutually dependent and thus cannot be solved in isolation. We characterize this challenge as stance entanglement, a form of decision complexity distinct from execution complexity. To address it, we propose Multi-Agent Fictitious Play (MAFP), a novel MAS paradigm that represents stakeholder stances as agents and formulates decision-making as an equilibrium-seeking process. Built on the game-theoretic principle of fictitious play, MAFP iteratively updates each agent's decision by best responding to the empirical mixture of other agents' past decisions. This enables agents to expose and address one another's weaknesses, progressively improving decision quality and robustness. We evaluate MAFP on challenging decision-making tasks that test the capability of deciding strategies for competitive scenarios prior to acting. MAFP outperforms both single-round and multi-round baselines on two complementary metrics, tournament strength and robustness, demonstrating its effectiveness in addressing stance entanglement.", "authors": [ "Leyang Shen", "Yang Zhang", "Xiaoyan Zhao", "Chun Kai Ling", "Tat-Seng Chua" ], "categories": [ "cs.CL", "cs.MA" ], "paper_id": "http://arxiv.org/abs/2606.19308v1", "abstract_url": "https://arxiv.org/abs/2606.19308v1", "pdf_url": "https://arxiv.org/pdf/2606.19308v1", "published_at": "2026-06-17T17:31:06+00:00", "updated_at": "2026-06-17T17:31:06+00:00", "source": "arxiv", "date_label": "Published", "analysis": null, "tags": [ "评测", "方法" ], "topics": [ "LLM", "Language Model" ], "doi": null, "arxiv_id": "2606.19308", "source_variants": [ "arxiv" ], "source_urls": { "arxiv": "https://arxiv.org/abs/2606.19308v1" }, "relevance_score": 185, "match_reasons": [ "title matched \"language model\"", "title matched \"large language model\"", "title matched \"agent\"", "summary matched \"LLM\"", "summary matched \"reasoning\"", "has PDF", "has rich summary", "has complete metadata" ], "feedback_status": null, "feedback_note": null, "feedback_next_action": null, "feedback_due_date": null, "feedback_snoozed_until": null, "feedback_review_interval_days": null, "canonical_id": "arxiv:2606.19308" }, { "title": "IndicContextEval: A Benchmark for Evaluating Context Utilisation in Audio Large Language Models Across 8 Indic Languages", "summary": "AudioLLMs enable speech recognition conditioned on textual prompts such as domain descriptions or entity lists. However, it remains unclear whether these models genuinely utilise such context or rely on parametric knowledge learned during pretraining. Existing benchmarks cannot answer this question because they evaluate transcription under fixed prompting conditions and rarely include explicit contextual inputs. We introduce IndicContextEval, a 56-hour multilingual benchmark of natural speech from 555 speakers across 8 Indian languages and 23 professional domains. We design a 7-level prompting framework that progressively introduces contextual signals, including metadata, natural-language descriptions, entity lists in English and native script, and adversarial prompts with incorrect entities. Evaluating five models reveals substantial differences in context utilisation behaviour, highlighting the need for explicit evaluation of contextual grounding in AudioLLMs.", "authors": [ "Sakshi Joshi", "Dhruv Subhash Rathi", "Sanskar Singh", "Eldho Ittan George", "R J Hari", "Kaushal Bhogale", "Mitesh M. Khapra" ], "categories": [ "eess.AS", "cs.CL" ], "paper_id": "http://arxiv.org/abs/2606.19157v1", "abstract_url": "https://arxiv.org/abs/2606.19157v1", "pdf_url": "https://arxiv.org/pdf/2606.19157v1", "published_at": "2026-06-17T14:59:37+00:00", "updated_at": "2026-06-17T14:59:37+00:00", "source": "arxiv", "date_label": "Published", "analysis": null, "tags": [ "评测", "方法" ], "topics": [ "LLM", "Language Model" ], "doi": null, "arxiv_id": "2606.19157", "source_variants": [ "arxiv" ], "source_urls": { "arxiv": "https://arxiv.org/abs/2606.19157v1" }, "relevance_score": 182, "match_reasons": [ "title matched \"language model\"", "title matched \"large language model\"", "title matched \"benchmark\"", "summary matched \"LLM\"", "summary matched \"evaluation\"", "has PDF", "has rich summary", "has complete metadata" ], "feedback_status": null, "feedback_note": null, "feedback_next_action": null, "feedback_due_date": null, "feedback_snoozed_until": null, "feedback_review_interval_days": null, "canonical_id": "arxiv:2606.19157" }, { "title": "A Technical Taxonomy of LLM Agent Communication Protocols", "summary": "As large language models (LLMs) advance and multi-agent systems aim to overcome the limits of standalone agents, robust communication protocols are becoming essential infrastructure for distributed agent networks. Nonetheless, the fragmented protocol landscape presents a significant interoperability challenge. This study develops a technical taxonomy to classify and analyze LLM agent communication protocols. Following an established iterative method, we defined the taxonomy's purpose, meta-characteristic, and ending conditions, then performed five iterations, three empirical-to-conceptual and two conceptual-to-empirical, on nine actively maintained open-source protocols with demonstrable adoption. The taxonomy comprises five dimensions: counterparty, payload, interaction state, discovery mechanism, and schema flexibility. Classification reveals recurring architectural patterns: all sampled agent-to-agent protocols combine hybrid payloads with session-state persistence; most protocols support multiple predefined schemas, and two negotiate schemas at runtime, indicating a trend toward schema flexibility; decentralized discovery remains rare. Analysis suggests short-term convergence pressure toward protocols unifying agent-to-agent and agent-to-context (tool and data) communication. Long-term, however, no single protocol is likely to maximize versatility, efficiency, and portability simultaneously. The field will more likely evolve toward a federated, layered protocol stack. The framework guides protocol selection and highlights open research gaps such as privacy and policy enforcement.}", "authors": [ "Linus Sander", "Habtom Kahsay Gidey", "Alexander Lenz", "Alois Knoll" ], "categories": [ "cs.MA", "cs.AI", "cs.NI" ], "paper_id": "http://arxiv.org/abs/2606.19135v1", "abstract_url": "https://arxiv.org/abs/2606.19135v1", "pdf_url": "https://arxiv.org/pdf/2606.19135v1", "published_at": "2026-06-17T14:45:20+00:00", "updated_at": "2026-06-17T14:45:20+00:00", "source": "arxiv", "date_label": "Published", "analysis": null, "tags": [ "应用", "方法" ], "topics": [ "LLM", "Language Model" ], "doi": null, "arxiv_id": "2606.19135", "source_variants": [ "arxiv" ], "source_urls": { "arxiv": "https://arxiv.org/abs/2606.19135v1" }, "relevance_score": 160, "match_reasons": [ "title matched \"LLM\"", "title matched \"agent\"", "summary matched \"language model\"", "summary matched \"large language model\"", "summary matched \"RAG\"", "has PDF", "has rich summary", "has complete metadata", "summary matched \"policy enforcement\"" ], "feedback_status": null, "feedback_note": null, "feedback_next_action": null, "feedback_due_date": null, "feedback_snoozed_until": null, "feedback_review_interval_days": null, "canonical_id": "arxiv:2606.19135" }, { "title": "Quantifying and Auditing LLM Evaluation via Positive--Unlabeled Learning", "summary": "Large Language Models (LLMs) are increasingly used as judges for scalable evaluation, yet such LLM--as--a--Judge systems exhibit systematic biases that are decoupled from semantic quality, most notably verbosity bias. Meanwhile, human supervision is costly and typically selective, yielding reliable positive judgments but leaving most outputs unlabelled and potentially mixed in quality. We formulate LLM evaluation under selective human supervision as a positive--unlabelled learning problem and propose a geometric auditing framework based on Partial Optimal Transport. By aligning a small set of human--verified positives with a reliable subset of unlabelled outputs in a fixed embedding space, our method identifies human--consistent preferences and corrects biased judges without retraining. Experiments demonstrate improved alignment with human preferences, increased robustness to presentation biases, and interpretable confidence estimates, offering a scalable and statistically grounded alternative to existing LLM--as--a--judge pipelines.", "authors": [ "Zilong Zhang", "Yi-Ting Hung", "Lei Ding", "Chi-Kuang Yeh" ], "categories": [ "stat.ML", "cs.LG", "stat.CO", "stat.ME" ], "paper_id": "http://arxiv.org/abs/2606.19057v1", "abstract_url": "https://arxiv.org/abs/2606.19057v1", "pdf_url": "https://arxiv.org/pdf/2606.19057v1", "published_at": "2026-06-17T13:26:04+00:00", "updated_at": "2026-06-17T13:26:04+00:00", "source": "arxiv", "date_label": "Published", "analysis": null, "tags": [ "评测", "方法" ], "topics": [ "LLM", "Language Model" ], "doi": null, "arxiv_id": "2606.19057", "source_variants": [ "arxiv" ], "source_urls": { "arxiv": "https://arxiv.org/abs/2606.19057v1" }, "relevance_score": 159, "match_reasons": [ "title matched \"LLM\"", "title matched \"evaluation\"", "summary matched \"language model\"", "summary matched \"large language model\"", "summary matched \"alignment\"", "has PDF", "has rich summary", "has complete metadata" ], "feedback_status": null, "feedback_note": null, "feedback_next_action": null, "feedback_due_date": null, "feedback_snoozed_until": null, "feedback_review_interval_days": null, "canonical_id": "arxiv:2606.19057" }, { "title": "Trade-offs in Medical LLM Adaptation: An Empirical Study in French QA", "summary": "The development of large language models (LLMs) has led to an increased focus on their adaptation to specialized domains and languages, yet the effectiveness of domain adaptation strategies remains unclear. We present a study of medical domain adaptation using French medical question-answering (QA) as a case study. We compare continual pretraining (CPT), supervised fine-tuning (SFT), and their combination across three model families, multiple sizes, and three initialization types, explicitly disentangling adaptation effects from base model choice. We evaluate both multiple-choice (MCQA) and open-ended QA (OEQA) under greedy and constrained decoding using automatic metrics and LLM-as-a-Judge evaluation. For MCQA, CPT+SFT most often achieves the best scores, but gains over SFT are small and frequently not statistically significant, making SFT a strong and cost-effective default. For OEQA, CPT consistently improves overlap-based metrics, while SFT often degrades generation quality; instruction tuning and CPT+SFT are preferred by LLM-based evaluation. Cross-lingual experiments further show effective transfer from French adaptation to English benchmarks. Overall, we provide practical guidelines for selecting adaptation strategies under computational constraints.", "authors": [ "Ikram Belmadani", "Oumaima El Khettari", "Carlos Ramisch", "Frederic Bechet", "Richard Dufour", "Benoit Favre" ], "categories": [ "cs.CL", "cs.AI" ], "paper_id": "http://arxiv.org/abs/2606.19266v1", "abstract_url": "https://arxiv.org/abs/2606.19266v1", "pdf_url": "https://arxiv.org/pdf/2606.19266v1", "published_at": "2026-06-17T16:42:22+00:00", "updated_at": "2026-06-17T16:42:22+00:00", "source": "arxiv", "date_label": "Published", "analysis": null, "tags": [ "评测", "应用", "方法" ], "topics": [ "LLM", "Language Model" ], "doi": null, "arxiv_id": "2606.19266", "source_variants": [ "arxiv" ], "source_urls": { "arxiv": "https://arxiv.org/abs/2606.19266v1" }, "relevance_score": 158, "match_reasons": [ "title matched \"LLM\"", "summary matched \"language model\"", "summary matched \"large language model\"", "summary matched \"instruction tuning\"", "summary matched \"benchmark\"", "summary matched \"evaluation\"", "has PDF", "has rich summary", "has complete metadata" ], "feedback_status": null, "feedback_note": null, "feedback_next_action": null, "feedback_due_date": null, "feedback_snoozed_until": null, "feedback_review_interval_days": null, "canonical_id": "arxiv:2606.19266" }, { "title": "X+Slides: Benchmarking Audience-Conditioned Slide Generation", "summary": "Automatically generating slide decks from source documents is an important application of large language models (LLMs). Existing benchmarks primarily assess slide completeness and technical depth, while overlooking the target audience as a critical real-world factor. For instance, specialists demand rigorous proofs, whereas decision-makers prioritize actionable conclusions. To bridge this gap, we introduce X+Slides, a benchmark specifically designed for audience-conditioned slide generation. Built on a diverse corpus spanning 113 topics and seven presentation scenes, X+Slides employs a dynamic evaluation framework constructed from 8,133 deduplicated, source-grounded probes. By assigning audience-specific utility weights to the same source-grounded probes, X+Slides reports four complementary metrics: Audience Coverage measures how much audience-essential information is conveyed, Domain-wise Coverage shows which information types are covered, Efficiency measures delivered utility per unit of attention cost, and Correctness verifies whether slide claims are supported by the source. Experiments on DeepPresenter, SlideTailor, and NotebookLM show that current systems can recover a substantial but still incomplete part of audience-essential information: at $τ_A=0.7$, DeepPresenter reaches a best Audience Coverage of 0.714, SlideTailor reaches 0.594, and the NotebookLM ablation reaches 0.853 while showing clear grounding differences. These results indicate that visual quality and broad topic coverage should not be treated as evidence support without source-grounded evaluation.", "authors": [ "Haodong Chen", "Xuanhe Zhou", "Wei Zhou", "Xinyue Shao", "Yanbing Zhu", "Bo Wang", "Jiawei Hong", "Anya Jia", "Fan Wu" ], "categories": [ "cs.AI" ], "paper_id": "http://arxiv.org/abs/2606.19256v1", "abstract_url": "https://arxiv.org/abs/2606.19256v1", "pdf_url": "https://arxiv.org/pdf/2606.19256v1", "published_at": "2026-06-17T16:30:26+00:00", "updated_at": "2026-06-17T16:30:26+00:00", "source": "arxiv", "date_label": "Published", "analysis": null, "tags": [ "评测", "数据", "应用", "方法" ], "topics": [ "LLM", "Language Model" ], "doi": null, "arxiv_id": "2606.19256", "source_variants": [ "arxiv" ], "source_urls": { "arxiv": "https://arxiv.org/abs/2606.19256v1" }, "relevance_score": 158, "match_reasons": [ "title matched \"benchmark\"", "summary matched \"language model\"", "summary matched \"large language model\"", "summary matched \"LLM\"", "summary matched \"RAG\"", "summary matched \"evaluation\"", "has PDF", "has rich summary", "has complete metadata" ], "feedback_status": null, "feedback_note": null, "feedback_next_action": null, "feedback_due_date": null, "feedback_snoozed_until": null, "feedback_review_interval_days": null, "canonical_id": "arxiv:2606.19256" }, { "title": "CAPRA: Scaling Feedback on Software Architecture Deliverables with a Multi-Agent LLM System", "summary": "Automated assessment in software engineering education has advanced significantly for code grading and essay scoring. However, reviewing software architecture deliverables, which requires analyzing structural completeness and requirements traceability, has not yet been fully automated. Applying Large Language Models (LLMs) to this task requires robust architectures to ensure technical feedback is accurate and reliable for students. This paper presents CAPRA (Configurable Architecture Proficiency Report Assessment), a multi-agent LLM system that analyzes software architecture deliverables to generate personalized, template-compliant LaTeX feedback. As a core design choice, CAPRA coordinates multiple specialized agents and employs a Python-based microservice for multi-modal document extraction, utilizing PyMuPDF and vision-enabled LLMs (specifically gpt-4o) to parse text and UML diagrams. To ensure educational reliability and mitigate hallucinations, CAPRA introduces a deterministic Evidence Anchoring step using fuzzy matching via normalized Levenshtein distance, along with a ConsistencyManager agent that cross-verifies, deduplicates, and merges findings. System performance is assessed using a structured eight-criterion binary evaluation taxonomy covering: (i) extraction completeness, (ii) feature validation, (iii) issue grounding and severity detection, (iv) recommendation specificity and traceability, and (v) template and tone compliance. A preliminary empirical evaluation on 10 student reports shows that CAPRA satisfied 88.8% of the evaluated criteria under a strict two-rater aggregation rule, achieved moderate inter-rater agreement with human evaluators (kappa = 0.582), and processed each report in slightly over 4 minutes. While these results support the viability of LLM-supported architectural feedback, human oversight remains essential for subjective assessment dimensions.", "authors": [ "Marco Becattini", "Niccolò Caselli", "Matteo Minin", "Roberto Verdecchia", "Enrico Vicario" ], "categories": [ "cs.SE", "cs.AI" ], "paper_id": "http://arxiv.org/abs/2606.18976v1", "abstract_url": "https://arxiv.org/abs/2606.18976v1", "pdf_url": "https://arxiv.org/pdf/2606.18976v1", "published_at": "2026-06-17T12:00:21+00:00", "updated_at": "2026-06-17T12:00:21+00:00", "source": "arxiv", "date_label": "Published", "analysis": null, "tags": [ "评测", "方法" ], "topics": [ "LLM", "Language Model" ], "doi": null, "arxiv_id": "2606.18976", "source_variants": [ "arxiv" ], "source_urls": { "arxiv": "https://arxiv.org/abs/2606.18976v1" }, "relevance_score": 157, "match_reasons": [ "title matched \"LLM\"", "title matched \"agent\"", "summary matched \"language model\"", "summary matched \"large language model\"", "summary matched \"evaluation\"", "has PDF", "has rich summary", "has complete metadata" ], "feedback_status": null, "feedback_note": null, "feedback_next_action": null, "feedback_due_date": null, "feedback_snoozed_until": null, "feedback_review_interval_days": null, "canonical_id": "arxiv:2606.18976" }, { "title": "Beyond Tokenization: Direct Timestep Embedding and Contrastive Alignment for Time-Series Question Answering", "summary": "Recent advances in large language models (LLMs) have given rise to time-series question answering (TSQA), which formulates time-series analysis as natural-language question answering. However, directly feeding raw numerical series into LLMs suffers from a tokenization bottleneck: Byte Pair Encoding fragments continuous values into unstable tokens whose embeddings lack meaningful metric structure, resulting in the loss of magnitude, scale, and trend information. Prior methods use patch-based encoders that split the series into fixed windows, locking in one granularity that breaks patterns and hides exact timesteps, through a separate module that rarely transfers across datasets with different lengths or sampling rates. To address this challenge, we propose CADE (Contrastive Alignment with Direct Embedding), a novel framework for TSQA built upon two key components: direct timestep embedding and semantic alignment. The proposed framework maps each timestep directly into the LLM embedding space through a point-wise linear encoder and MLP projector, preserving exact index-level access while eliminating the need for patching and padding. To further bridge the semantic gap between time-series and language representations, we introduce a novel one-directional supervised contrastive loss that aligns time-series embeddings with frozen class-name text anchors. Experimental results on the public Time-MQA benchmark demonstrate that our framework consistently improves performance across six TSQA tasks, outperforming both open-source and proprietary LLM baselines.", "authors": [ "Yafeng Wu", "Huu Hiep Nguyen", "Thin Nguyen", "Hung Le" ], "categories": [ "cs.CL", "cs.AI" ], "paper_id": "http://arxiv.org/abs/2606.18986v1", "abstract_url": "https://arxiv.org/abs/2606.18986v1", "pdf_url": "https://arxiv.org/pdf/2606.18986v1", "published_at": "2026-06-17T12:07:23+00:00", "updated_at": "2026-06-17T12:07:23+00:00", "source": "arxiv", "date_label": "Published", "analysis": null, "tags": [ "评测", "数据", "方法" ], "topics": [ "LLM", "Language Model" ], "doi": null, "arxiv_id": "2606.18986", "source_variants": [ "arxiv" ], "source_urls": { "arxiv": "https://arxiv.org/abs/2606.18986v1" }, "relevance_score": 154, "match_reasons": [ "title matched \"alignment\"", "summary matched \"language model\"", "summary matched \"large language model\"", "summary matched \"LLM\"", "summary matched \"RAG\"", "summary matched \"benchmark\"", "has PDF", "has rich summary", "has complete metadata" ], "feedback_status": null, "feedback_note": null, "feedback_next_action": null, "feedback_due_date": null, "feedback_snoozed_until": null, "feedback_review_interval_days": null, "canonical_id": "arxiv:2606.18986" }, { "title": "Dango: A Strictly L1-Only Large Language Model for Studying Second Language Acquisition", "summary": "We introduce Dango, a 1.8B-parameter large language model designed for controlled studies of L1-to-L2 (Japanese-to-English) transfer in second language acquisition (SLA). While previous studies have explored SLA in language models, they have predominantly relied on smaller or non-decoder models, limiting their ability to generate open-ended text and reducing their suitability as practical L2 simulators. We identify a key challenge when scaling models to this size: L2 contamination within the \"monolingual\" pretraining corpus used for L1 acquisition. To address this, we propose a filtering method to reduce premature exposure to English while preserving realistic, minimal exposure. We then fine-tune the model on LLM-generated L2-learning lessons to simulate the L2 acquisition process. Our evaluations confirm that Dango develops human-like L2 production patterns, outperforming both unfiltered and standard multilingual baselines. We release the model, data, and code to facilitate reproducible computational SLA research and learner-facing applications.", "authors": [ "Shiho Matta", "Yin Jou Huang", "Fei Cheng", "Takashi Kodama", "Hirokazu Kiyomaru", "Yugo Murawaki" ], "categories": [ "cs.CL" ], "paper_id": "http://arxiv.org/abs/2606.19170v1", "abstract_url": "https://arxiv.org/abs/2606.19170v1", "pdf_url": "https://arxiv.org/pdf/2606.19170v1", "published_at": "2026-06-17T15:13:19+00:00", "updated_at": "2026-06-17T15:13:19+00:00", "source": "arxiv", "date_label": "Published", "analysis": null, "tags": [ "评测", "数据", "应用", "方法" ], "topics": [ "LLM", "Language Model" ], "doi": null, "arxiv_id": "2606.19170", "source_variants": [ "arxiv" ], "source_urls": { "arxiv": "https://arxiv.org/abs/2606.19170v1" }, "relevance_score": 143, "match_reasons": [ "title matched \"language model\"", "title matched \"large language model\"", "summary matched \"LLM\"", "summary matched \"evaluation\"", "has PDF", "has rich summary", "has complete metadata" ], "feedback_status": null, "feedback_note": null, "feedback_next_action": null, "feedback_due_date": null, "feedback_snoozed_until": null, "feedback_review_interval_days": null, "canonical_id": "arxiv:2606.19170" }, { "title": "G-IdiomAlign: A Gloss-Pivoted Benchmark for Cross-Lingual Idiom Alignment", "summary": "Idioms are difficult to transfer across languages due to their non-compositionality and weak surface-form grounding, making literal mappings unreliable. We present G-IdiomAlign, a gloss-pivoted benchmark where each idiom is anchored by an English gloss from Wiktionary. We further construct a high-confidence reference alignment set for reproducible evaluation. G-IdiomAlign supports two protocols: (1) a controlled Multiple-Choice Idiom Equivalence with typed distractors for error attribution; and (2) a Gloss-Contrastive Generation contrasting No-gloss and With-gloss inputs to isolate the effect of an explicit semantic pivot. Across diverse LLMs, a bias to literal translation is a dominant failure mode, especially when the target is a low-resource language. Glosses consistently improve Gloss-Contrastive Generation under an embedding-based semantic proxy, but performance remains modest, indicating substantial headroom in the open output space. Subsequent analysis on Qwen3-8B further suggests that cross-condition differences are concentrated more in attention heads than in layers, while better With-gloss generations coincide with stronger gloss anchoring.", "authors": [ "Fengying Ye", "Yanming Sun", "Runzhe Zhan", "Zheqi Zhang", "Lidia S. Chao", "Derek F. Wong" ], "categories": [ "cs.CL", "cs.AI" ], "paper_id": "http://arxiv.org/abs/2606.18989v1", "abstract_url": "https://arxiv.org/abs/2606.18989v1", "pdf_url": "https://arxiv.org/pdf/2606.18989v1", "published_at": "2026-06-17T12:09:00+00:00", "updated_at": "2026-06-17T12:09:00+00:00", "source": "arxiv", "date_label": "Published", "analysis": null, "tags": [ "评测", "方法" ], "topics": [ "LLM", "Benchmark" ], "doi": null, "arxiv_id": "2606.18989", "source_variants": [ "arxiv" ], "source_urls": { "arxiv": "https://arxiv.org/abs/2606.18989v1" }, "relevance_score": 140, "match_reasons": [ "title matched \"alignment\"", "title matched \"benchmark\"", "summary matched \"LLM\"", "summary matched \"evaluation\"", "has PDF", "has rich summary", "has complete metadata" ], "feedback_status": null, "feedback_note": null, "feedback_next_action": null, "feedback_due_date": null, "feedback_snoozed_until": null, "feedback_review_interval_days": null, "canonical_id": "arxiv:2606.18989" }, { "title": "Beyond Safe Data: Pretraining-Stage Alignment with Regular Safety Reflection", "summary": "To achieve deeper safety alignment for large language models (LLMs), recent efforts have studied how to push safety interventions earlier into the pretraining stage, primarily by filtering unsafe data or rewriting it into safer forms. We argue that pretraining-stage alignment should go beyond making the data safe: LLMs may compose seemingly benign knowledge and capabilities into unsafe behaviors. To this end, we propose Safety Reflection Pretraining, a pretraining-stage alignment method which regularly inserts short safety reflections into pretraining corpora to integrate self-monitoring directly into language modeling, establishing a foundational capability that is subsequently reinforced by compatible post-training. Our experiments with 1.7B models pretrained on FineWeb-Edu show that Safety Reflection Pretraining improves safety classification accuracy and substantially reduces the success rates of inference-stage and finetuning attacks. Complementary to our real-world experiments, we also introduce a fully controlled synthetic environment, MedSafetyWorld, with a clear definition of safety and a reasoning structure under which models can easily generalize unsafe behaviors from safe data. Ablations in MedSafetyWorld further demonstrate a clear advantage of Safety Reflection Pretraining in preventing models from acting on unsafe behaviors generalized from safe data, compared with data filtering and rewriting. Taken together, our findings suggest that pretraining alignment should not only make the training data safe, but also shape the behaviors that models are likely to acquire from safe data.", "authors": [ "Jinhan Li", "Kexian Tang", "Yihan Xu", "Zhuorui Ye", "Kaifeng Lyu" ], "categories": [ "cs.AI", "cs.LG" ], "paper_id": "http://arxiv.org/abs/2606.19168v1", "abstract_url": "https://arxiv.org/abs/2606.19168v1", "pdf_url": "https://arxiv.org/pdf/2606.19168v1", "published_at": "2026-06-17T15:11:43+00:00", "updated_at": "2026-06-17T15:11:43+00:00", "source": "arxiv", "date_label": "Published", "analysis": null, "tags": [ "方法" ], "topics": [ "LLM", "Language Model" ], "doi": null, "arxiv_id": "2606.19168", "source_variants": [ "arxiv" ], "source_urls": { "arxiv": "https://arxiv.org/abs/2606.19168v1" }, "relevance_score": 139, "match_reasons": [ "title matched \"alignment\"", "summary matched \"language model\"", "summary matched \"large language model\"", "summary matched \"LLM\"", "summary matched \"reasoning\"", "has PDF", "has rich summary", "has complete metadata" ], "feedback_status": null, "feedback_note": null, "feedback_next_action": null, "feedback_due_date": null, "feedback_snoozed_until": null, "feedback_review_interval_days": null, "canonical_id": "arxiv:2606.19168" }, { "title": "AdsMind: A Physics-Grounded Multi-Agent System for Self-Correcting Discovery of Adsorption Configurations on Heterogeneous Catalyst Surfaces", "summary": "Identifying the lowest-energy surface-adsorbate configuration is critical for modeling heterogeneous catalysis, yet exhaustive exploration with ab initio calculations is computationally prohibitive. Machine-learning force fields (MLFFs) accelerate structural relaxation but leave the search over the vast configurational space a major bottleneck, and open-loop large language model (LLM) agents lack a physics-grounded feedback mechanism to correct erroneous initial guesses. We propose AdsMind (Adsorption configuration discovery with Machine intelligence and relaxation feedback), a closed-loop multi-agent framework that enables autonomous error correction through MLFF relaxation feedback. Across four LLM backends, AdsMind achieves consistently high search reliability, with success rates of 100% and 98.8% on the benchmarks AA20 and OCD-GMAE62. Relative to its single-pass (1-Shot) ablation it reduces cross-backend energy dispersion, and it uses only 4.11 and 4.67 MLFF relaxations per case, respectively -- an approximately 14-fold reduction over heuristic enumeration baselines. Density functional theory (DFT) validation using VASP/PBE on six representative AA20 systems shows that the reported open-loop Adsorb-Agent outputs exhibit qualitative adsorption-energy sign errors for molecular adsorbates, whereas AdsMind preserves the correct sign in all tested cases with closer quantitative agreement. AdsMind thus delivers reliability, self-reflection, and interpretability simultaneously, supporting more DFT-informed autonomous chemistry workflows.", "authors": [ "Zongmin Zhang", "Yuyang Lou", "Bowen Zhang", "Junwu Chen", "Ryo Kuroki", "Xuan Vu Nguyen", "Edvin Fako", "Lixue Cheng", "Philippe Schwaller" ], "categories": [ "cond-mat.mtrl-sci", "cs.AI" ], "paper_id": "http://arxiv.org/abs/2606.19152v1", "abstract_url": "https://arxiv.org/abs/2606.19152v1", "pdf_url": "https://arxiv.org/pdf/2606.19152v1", "published_at": "2026-06-17T14:57:16+00:00", "updated_at": "2026-06-17T14:57:16+00:00", "source": "arxiv", "date_label": "Published", "analysis": null, "tags": [ "评测", "应用", "方法" ], "topics": [ "LLM", "Language Model" ], "doi": null, "arxiv_id": "2606.19152", "source_variants": [ "arxiv" ], "source_urls": { "arxiv": "https://arxiv.org/abs/2606.19152v1" }, "relevance_score": 138, "match_reasons": [ "title matched \"agent\"", "summary matched \"language model\"", "summary matched \"large language model\"", "summary matched \"LLM\"", "summary matched \"benchmark\"", "has PDF", "has rich summary", "has complete metadata" ], "feedback_status": null, "feedback_note": null, "feedback_next_action": null, "feedback_due_date": null, "feedback_snoozed_until": null, "feedback_review_interval_days": null, "canonical_id": "arxiv:2606.19152" }, { "title": "OpenAnt: LLM-Powered Vulnerability Discovery Through Code Decomposition, Adversarial Verification, and Dynamic Testing", "summary": "Automated vulnerability discovery in large codebases remains challenging: traditional static analysis produces high false-positive rates, while dynamic approaches such as fuzzing require substantial infrastructure and often target narrow classes of bugs. Recent advances in large language models (LLMs) enable semantic reasoning about program behavior, but applying LLMs to repository-scale security analysis introduces challenges related to context management, cost, and verification. We present OpenAnt, an open-source vulnerability discovery system that integrates static program analysis with LLM-based reasoning in a multi-stage pipeline. OpenAnt introduces three key techniques. First, codebases are decomposed into self-contained analysis units filtered by reachability from external entry points, reducing the analysis surface by up to 97% while preserving attack-relevant code. Second, candidate vulnerabilities undergo adversarial verification through constrained attacker simulation, where the model evaluates exploitability under realistic attacker capabilities. Third, findings are validated through dynamic verification, in which exploit environments are generated automatically, executed in sandboxed containers, and discarded after use. Evaluation on widely used open-source projects including OpenSSL, WordPress, and Flowise shows that this architecture can identify previously unknown vulnerabilities while maintaining manageable analysis cost and substantially reducing false positives. Our results suggest that closed-loop vulnerability discovery pipelines, combining semantic reasoning with exploit validation, provide a practical path toward scalable automated security analysis. OpenAnt is released as open source under the Apache 2.0 license at https://github.com/knostic/OpenAnt.", "authors": [ "Nahum Korda", "Gadi Evron" ], "categories": [ "cs.CR", "cs.LG" ], "paper_id": "http://arxiv.org/abs/2606.19149v1", "abstract_url": "https://arxiv.org/abs/2606.19149v1", "pdf_url": "https://arxiv.org/pdf/2606.19149v1", "published_at": "2026-06-17T14:56:04+00:00", "updated_at": "2026-06-17T14:56:04+00:00", "source": "arxiv", "date_label": "Published", "analysis": null, "tags": [ "评测", "方法" ], "topics": [ "LLM", "Language Model" ], "doi": null, "arxiv_id": "2606.19149", "source_variants": [ "arxiv" ], "source_urls": { "arxiv": "https://arxiv.org/abs/2606.19149v1" }, "relevance_score": 138, "match_reasons": [ "title matched \"LLM\"", "summary matched \"language model\"", "summary matched \"large language model\"", "summary matched \"reasoning\"", "summary matched \"evaluation\"", "has PDF", "has rich summary", "has complete metadata" ], "feedback_status": null, "feedback_note": null, "feedback_next_action": null, "feedback_due_date": null, "feedback_snoozed_until": null, "feedback_review_interval_days": null, "canonical_id": "arxiv:2606.19149" }, { "title": "Seeing Before Reasoning: Decoupling Perception and Reasoning for Shortcut-Resilient Multimodal On-Policy Self-Distillation", "summary": "On-policy self-distillation (OPSD) trains a model on its own rollouts and uses a frozen copy to provide dense token-level targets conditioned on a reference target. This works well for LLM reasoning, but a direct extension to multimodal large language models (MLLMs) can create a shortcut: the privileged target may guide tokens mainly based on the text reference target rather than the image. We propose ViGOS, a visually grounded OPSD framework for MLLM post-training. The student first writes a visual description and then reasons toward the final answer. For valid rollouts, an image-only perception teacher supervises the description, while a privileged reasoning teacher supervises the reasoning and final answer on the same student prefix. A reference teacher is used only for invalid rollouts to recover the output format. Across general vision-language, expert reasoning, visual math, spatial grounding, and visual-language-prior benchmarks, ViGOS keeps the main benefits of OPSD and improves image-grounded behavior in shortcut-prone settings.", "authors": [ "Sihan Wang", "Xiyao Liu", "Lianqing Liu", "Zhi Han" ], "categories": [ "cs.LG", "cs.CV" ], "paper_id": "http://arxiv.org/abs/2606.19120v1", "abstract_url": "https://arxiv.org/abs/2606.19120v1", "pdf_url": "https://arxiv.org/pdf/2606.19120v1", "published_at": "2026-06-17T14:33:38+00:00", "updated_at": "2026-06-17T14:33:38+00:00", "source": "arxiv", "date_label": "Published", "analysis": null, "tags": [ "评测", "方法" ], "topics": [ "LLM", "Language Model" ], "doi": null, "arxiv_id": "2606.19120", "source_variants": [ "arxiv" ], "source_urls": { "arxiv": "https://arxiv.org/abs/2606.19120v1" }, "relevance_score": 138, "match_reasons": [ "title matched \"reasoning\"", "summary matched \"language model\"", "summary matched \"large language model\"", "summary matched \"LLM\"", "summary matched \"benchmark\"", "has PDF", "has rich summary", "has complete metadata" ], "feedback_status": null, "feedback_note": null, "feedback_next_action": null, "feedback_due_date": null, "feedback_snoozed_until": null, "feedback_review_interval_days": null, "canonical_id": "arxiv:2606.19120" }, { "title": "Diffusion-Proof: Recipe for Formal Theorem Proving Beyond Auto-Regressive Generation", "summary": "Enhancing the formal math reasoning capabilities of Large Language Models (LLMs) has become a key focus in both mathematical and computer science communities in recent years. While significant progress has been made in using state-of-the-art Auto-Regressive (AR) LLMs for formal theorem proving, these models suffer from inherent limitations. Their next-token prediction generation methods may yield suboptimal performance due to the challenges of long-range coherence and the compounding of errors over long sequences. Recent advancements in diffusion LLMs (dLLMs), which generate text through iterative denoising of a multi-token block, offer a promising alternative. However, the application of dLLMs to formal mathematics, where maintaining long-range coherence is critical, remains largely understudied. To address the challenges above, we propose **Diffusion-Proof**, to the best of our knowledge, the first framework to train and apply dLLMs for formal theorem proving. Our frameworks contain training and inference methods for two models. The first one is *dLLM-Prover-7B*, which performs whole-proof writing with long-range coherent tactic usage. The second one is *dLLM-Corrector-7B*, which is a novel large block diffusion-based correction model. It leverages the in-filling capabilities of dLLMs to perform local proof correction using bi-directional information. Extensive experiments demonstrate that **Diffusion-Proof** relatively significantly outperforms the AR LLM baseline trained under the same dataset. **Diffusion-Proof** achieves an absolute improvement of **1.61%** on ProofNet-Test and **6.14%** on MiniF2F-Test benchmarks compare to the baseline. Notably, **Diffusion-Proof** successfully resolves one IMO problem that more advanced thinking model DeepSeek-Prover-V2-7B could not solve, showcasing the unique advantage of dLLMs in formal theorem proving.", "authors": [ "Ruida Wang", "Rui Pan", "Pengcheng Wang", "Shizhe Diao", "Tong Zhang" ], "categories": [ "cs.LG" ], "paper_id": "http://arxiv.org/abs/2606.19315v1", "abstract_url": "https://arxiv.org/abs/2606.19315v1", "pdf_url": "https://arxiv.org/pdf/2606.19315v1", "published_at": "2026-06-17T17:38:32+00:00", "updated_at": "2026-06-17T17:38:32+00:00", "source": "arxiv", "date_label": "Published", "analysis": null, "tags": [ "评测", "数据", "应用", "方法" ], "topics": [ "LLM", "Language Model" ], "doi": null, "arxiv_id": "2606.19315", "source_variants": [ "arxiv" ], "source_urls": { "arxiv": "https://arxiv.org/abs/2606.19315v1" }, "relevance_score": 137, "match_reasons": [ "summary matched \"language model\"", "summary matched \"large language model\"", "summary matched \"LLM\"", "summary matched \"reasoning\"", "summary matched \"RAG\"", "summary matched \"benchmark\"", "has PDF", "has rich summary", "has complete metadata" ], "feedback_status": null, "feedback_note": null, "feedback_next_action": null, "feedback_due_date": null, "feedback_snoozed_until": null, "feedback_review_interval_days": null, "canonical_id": "arxiv:2606.19315" } ] }, { "name": "Agent Runtime Security", "key_points": [ "《CodeSentinel: A Three-Layer Defense Against Indirect Prompt Injection in Code Contexts》〔方法〕：Code large language models increasingly retrieve external code context from repositories, documentation, issue threads, and coding-agent environments, creating…" ], "sort_by": "hybrid", "papers": [ { "title": "CodeSentinel: A Three-Layer Defense Against Indirect Prompt Injection in Code Contexts", "summary": "Code large language models increasingly retrieve external code context from repositories, documentation, issue threads, and coding-agent environments, creating an indirect prompt-injection surface where attackers hide instructions in comments, strings, identifiers, or decoy code. We propose CodeSentinel, a three-layer inference-time sanitizer. It uses Tree-sitter to extract high-risk model-facing CST nodes, then combines syntax-guided pre-filtering, CST-guided Dynamic Min-K\\% scoring, and node perturbation analysis to detect adversarial and natural-looking semantic triggers. Detected nodes are removed or neutralized before reaching the downstream Code LLM. Across six recent attack families, \\CodeSentinel achieves 0.80 average node-level F1, outperforming CodeGarrison, DePA, and KillBadCode.", "authors": [ "Po-Han Cheng", "Chia-Mu Yu", "Ying-Dar Lin", "Yu-Sung Wu", "Wei-Bin Lee" ], "categories": [ "cs.CR" ], "paper_id": "http://arxiv.org/abs/2606.19235v1", "abstract_url": "https://arxiv.org/abs/2606.19235v1", "pdf_url": "https://arxiv.org/pdf/2606.19235v1", "published_at": "2026-06-17T16:12:50+00:00", "updated_at": "2026-06-17T16:12:50+00:00", "source": "arxiv", "date_label": "Published", "analysis": null, "tags": [ "方法" ], "topics": [ "LLM", "Language Model" ], "doi": null, "arxiv_id": "2606.19235", "source_variants": [ "arxiv" ], "source_urls": { "arxiv": "https://arxiv.org/abs/2606.19235v1" }, "relevance_score": 108, "match_reasons": [ "title matched \"prompt injection\"", "title matched \"indirect prompt injection\"", "has PDF", "has rich summary", "has complete metadata" ], "feedback_status": null, "feedback_note": null, "feedback_next_action": null, "feedback_due_date": null, "feedback_snoozed_until": null, "feedback_review_interval_days": null, "canonical_id": "arxiv:2606.19235" } ] }, { "name": "Terminal and SWE Agents", "key_points": [ "《Data Intelligence Agents: Interpreting, Modeling, and Querying Enterprise Data via Autonomous Coding Agents》〔评测 / 应用 / 方法〕：Production data integration is bottlenecked by repeated, lossy handoffs between data owners, engineers, and analysts who must collaboratively discover, structu…" ], "sort_by": "hybrid", "papers": [ { "title": "Data Intelligence Agents: Interpreting, Modeling, and Querying Enterprise Data via Autonomous Coding Agents", "summary": "Production data integration is bottlenecked by repeated, lossy handoffs between data owners, engineers, and analysts who must collaboratively discover, structure, and query enterprise data. We present Data Intelligence Agents (DIA), a system of three agents (Data Interpreter, Schema Creator, and Query Generator) that compresses this workflow by treating autonomous coding agents (ACAs) as a first-class abstraction: rather than emitting text, the agents generate, execute, validate, and repair concrete artifacts, draw on a shared memory for experience reuse, and surface each for review by domain experts. DIA is deployed in production for enterprise customers. We study the Query Generator in depth and evaluate it in fully autonomous mode across seven SQL benchmarks spanning four task categories and four dialects. It matches or surpasses the best published results on all seven, demonstrating that an architecture grounded in execution, built on ACAs and a shared memory, generalizes across the data intelligence workload with adaptation confined to natural-language instructions.", "authors": [ "Anoushka Vyas", "Aarushi Dhanuka", "Sina Khoshfetrat Pakazad", "Henrik Ohlsson" ], "categories": [ "cs.MA", "cs.AI", "cs.DB" ], "paper_id": "http://arxiv.org/abs/2606.19319v1", "abstract_url": "https://arxiv.org/abs/2606.19319v1", "pdf_url": "https://arxiv.org/pdf/2606.19319v1", "published_at": "2026-06-17T17:45:32+00:00", "updated_at": "2026-06-17T17:45:32+00:00", "source": "arxiv", "date_label": "Published", "analysis": null, "tags": [ "评测", "应用", "方法" ], "topics": [ "Benchmark", "Agent" ], "doi": null, "arxiv_id": "2606.19319", "source_variants": [ "arxiv" ], "source_urls": { "arxiv": "https://arxiv.org/abs/2606.19319v1" }, "relevance_score": 69, "match_reasons": [ "title matched \"coding agent\"", "has PDF", "has rich summary", "has complete metadata" ], "feedback_status": null, "feedback_note": null, "feedback_next_action": null, "feedback_due_date": null, "feedback_snoozed_until": null, "feedback_review_interval_days": null, "canonical_id": "arxiv:2606.19319" } ] } ] }