arXiv cs.AI - 2026-06-17

342 items collected.

1. Beyond Parallel Sampling: Diverse Query Initialization for Agentic Search

Author: Sidhaarth Murali, Jo~ao Coelho, Jingjie Ning, Jo~ao Magalh~aes, Bruno Martins, Chenyan Xiong
Published: 6/17/2026, 4:00:00 AM
Categories: cs.AI, cs.IR

arXiv:2606.17209v1 Announce Type: new Abstract: Test-time scaling for agentic search typically increases depth (i.e., more turns and tokens per trajectory) or breadth (i.e., more parallel rollouts). Here we focus on breadth scaling, showing that standard parallel sampling yields diminishing returns,...

arXiv cs.AI - 2026-06-17 ​

1. Beyond Parallel Sampling: Diverse Query Initialization for Agentic Search ​

2. When Rules Learn: A Self-Evolving Agent for Legal Case Retrieval ​

3. SkillChain-Gym: A Benchmark for Reskilling-Aware Production-Inventory Control under Disruptions ​

4. Skill-Constrained Model Predictive Control for Resilient Manufacturing Supply Chains ​

5. Nothing from Something: Can a Language Model Discover 0? ​

6. Quantifying Consistency in LLM Logical Reasoning via Structural Uncertainty ​

7. MemTrace: Probing What Final Accuracy Misses in Long-Term Memory ​

8. SpeechDx: A Multi-Task Benchmark for Clinical Speech AI ​

9. Distributed General-Purpose Agent Networks: Architecture, Key Mechanisms, and Prototypes ​

10. Treatment Response Optimized Clinical Decision Support AI System via Digital Twin Simulation ​

11. Incumbent Advantage: Brand Bias and Cognitive Manipulation Dynamics in LLM Recommendation Systems ​

12. A Machine-Learned Comorbidity Index ​

13. MapSatisfyBench: Benchmarking Satisfaction-Aware Map Agents through Behavior-Grounded Implicit Decision Factors ​

14. Dissecting model behavior through agent trajectories ​

15. Can LLMs Be CEOs? Benchmarking Strategic Resource Reallocation with Multi-Role Agent Simulation ​

16. LLM-as-Judge in Education: A Curriculum-Grounded Marking Pipeline ​

17. SEAGym: An Evaluation Environment for Self-Evolving LLM Agents ​

18. DeepInsight: A Unified Evaluation Infrastructure Across the Physical AI Stack ​

19. Surrogate Assisted Pedestrian Protection Design via a Foundation Model Orchestrated Workflow ​

20. Closing the Feedback Loop: From Experience Extraction to Insight Governance in Verbal Reinforcement Learning ​

21. Brick-DICL: Dynamic In-Context Learning for Automated Brick Schema Classification ​

22. FinAcumen: Financial Multimodal Reasoning via Self-Evolving Experience Memory Harness ​

23. Beyond Domains: Reusing Web Skills via Transferable Interaction Patterns ​

24. From Brewing to Resolution: Tracing the Internal Lifecycle of Code Reasoning in LLMs ​

25. Using Cognitive Models to Improve Language Model Simulation of Human Persuasion Games ​

26. FllumaOne: A Code-Native Multimodal CAD Dataset with Executable Programs and Kernel-Validated Feature Histories ​

27. EComAgentBench: Benchmarking Shopping Agents on Long-Horizon Tasks with Distributed Hidden Intent ​

28. LongWebBench: Evaluating Structural and Functional Webpage Generation in Long-Horizon Settings ​

29. Shattering the Autoregressive Curse: Dynamic Epistemic Entropy Orchestrated Erasable Reinforcement Learning for LLMs ​

30. DecoSearch: Complexity-Aware Routing and Plan-Level Repair for Text-to-SQL ​

31. WallZero: Mastering the Game of WallGo with Strategic Analysis ​

32. A homotopy-type-theoretic generalization of neurosymbolic inference ​

33. FlowRAG: Synergizing Explicit Reasoning via Frequency-Aware Multi-Granularity Graph Flow ​

34. StepGuard: Guarding Web Navigation via Single-Step Calibration ​

35. Structural Preservation and the Logical Expressiveness of Graph Neural Networks ​

36. MathVis-Fine: Aligning Visual Supervision with Necessity via Progressive Dependency-Guided Training for Multimodal Mathematical Reasoning ​

37. Learn to Quantify Social Interaction with Constraints for Pedestrian Walking ​

38. DiagFlowBench: Evaluating How Language Models Handle Off-Procedure Inputs in Grounded Diagnostic Dialogue ​

39. PreAct: Computer-Using Agents that Get Faster on Repeated Tasks ​

40. How Inference Compute Shapes Frontier LLM Evaluation ​

41. Small Initialization Matters for Large Language Models ​

42. MoCo-AIS: A Contrastive Learning Framework for Similarity Computation of Vessel Trajectories ​

43. STAR: SpatioTemporal Adaptive Reward Allocation for Text-to-Image RL Post-Training ​

44. LLM Consumer Behavior Theory: Foundations of a Novel Research Field ​

45. LegalHalluLens: Typed Hallucination Auditing and Calibrated Multi-Agent Debate for Trustworthy Legal AI ​

46. ProvenanceGuard: Source-Aware Factuality Verification for MCP-Based LLM Agents ​

47. PseudoBench: Measuring How Agentic Auto-Research Fuels Pseudoscience ​

48. Agentic AI-based Framework for Mitigating Premature Diagnostic Handoff and Silent Hallucination in Healthcare Applications ​

49. A Unified Framework for Context-Aware and Relation-Aware Graph Retrieval-Augmented Generation ​

50. IsabeLLM: Automated Theorem Proving Applied to Formally Verifying Consensus ​

51. Trust the Right Teacher: Quality-Aware Self-Distillation for GUI Grounding ​

52. First Proof Second Batch ​

53. Knowledge Reutilization in Meta-Reinforcement Learning ​

54. Your AI Travel Agent Would Book You a Bullfight: An Agentic Benchmark for Implicit Animal Welfare in Frontier AI Models ​

55. Memory as a Wasting Asset: Pricing Flash Endurance for Embodied Agents, and the Limits of Doing So ​

56. WEQA: Wearable hEalth Question Answering with Query-Adaptive Agentic Reasoning ​

57. Learning Cardiac Electrophysiology Digital Twins Through Agentic Discovery of Hybrid Structure ​

58. DRFLOW: A Deep Research Benchmark for Personalized Workflow Prediction ​

59. The Stanford EDGAR Filings Dataset: Reconstructing U.S. Corporate and Financial Disclosures into Layout-Faithful and Token-Efficient Pretraining Data ​

60. Fixed-Point Reasoners: Stable and Adaptive Deep Looped Transformers ​

61. EvolveNav: Proactive Preflection and Self-Evolving Memory for Zero-Shot Object Goal Navigation ​

62. Correct When Paired, Wrong When Split: Decoupling and Editing Modality-Specific Neurons in MLLMs ​

63. Towards Distributed Inference of LLMs on a P2P Network ​

64. PIVOT: Bridging Black-Scholes Implied-Volatility and Price Objectives via Differentiable J\"ackel Operator ​

65. KFTD: Koopman-Fourier Time-Differentiable Network for Continuous Ocean Spatiotemporal Forecasting ​

66. Extracting Semantics: LLM-Guided Automatic Population of Robot Ontology from URDF ​

67. Surveying GenAI-based Automation in Printed Circuit Board Design and Test ​

68. CMIP-Forge: An Agentic System that Retrieves, Computes, and Self-Reviews Climate Science ​

69. Comprehensive pKa Data Augmentation from Limited Real Data through an Engineered Models-Quantum Framework ​

70. HRDX: A Large-Scale Vector HD-Map Dataset ​

71. The Price of Anarchy in Disaggregated Inference ​

72. ParkingTransformer: LLM-Enhanced End-to-End Trajectory Planning for Autonomous Parking ​

73. ZIVARI-TLBO: A Zero-Cost Inter-Group Evaluated-Elite Relay Mechanism for Teaching-Learning-Based Optimization ​

74. ANEForge: Python for direct computation on the Apple Neural Engine ​

75. Software Delegation Contracts: Measuring Reviewability in AI Coding-Agent Work ​

76. Quantum Cinema: An Interactive Cinematic Exploration of Quantum Computing Hardware via Generative World Models ​

77. Prefill/Decode-Aware Evaluation of LLM Inference on Emerging AI Accelerators ​

78. Models Take Notes at Prefill: KV Cache Can Be Editable and Composable ​

79. Timestamp-Aware Spatio-Temporal Graph Contrastive Learning for Network Intrusion Detection ​

arXiv cs.AI - 2026-06-17

1. Beyond Parallel Sampling: Diverse Query Initialization for Agentic Search

2. When Rules Learn: A Self-Evolving Agent for Legal Case Retrieval

3. SkillChain-Gym: A Benchmark for Reskilling-Aware Production-Inventory Control under Disruptions

4. Skill-Constrained Model Predictive Control for Resilient Manufacturing Supply Chains

5. Nothing from Something: Can a Language Model Discover 0?

6. Quantifying Consistency in LLM Logical Reasoning via Structural Uncertainty

7. MemTrace: Probing What Final Accuracy Misses in Long-Term Memory

8. SpeechDx: A Multi-Task Benchmark for Clinical Speech AI

9. Distributed General-Purpose Agent Networks: Architecture, Key Mechanisms, and Prototypes

10. Treatment Response Optimized Clinical Decision Support AI System via Digital Twin Simulation

11. Incumbent Advantage: Brand Bias and Cognitive Manipulation Dynamics in LLM Recommendation Systems

12. A Machine-Learned Comorbidity Index

13. MapSatisfyBench: Benchmarking Satisfaction-Aware Map Agents through Behavior-Grounded Implicit Decision Factors

14. Dissecting model behavior through agent trajectories

15. Can LLMs Be CEOs? Benchmarking Strategic Resource Reallocation with Multi-Role Agent Simulation

16. LLM-as-Judge in Education: A Curriculum-Grounded Marking Pipeline

17. SEAGym: An Evaluation Environment for Self-Evolving LLM Agents

18. DeepInsight: A Unified Evaluation Infrastructure Across the Physical AI Stack

19. Surrogate Assisted Pedestrian Protection Design via a Foundation Model Orchestrated Workflow

20. Closing the Feedback Loop: From Experience Extraction to Insight Governance in Verbal Reinforcement Learning

21. Brick-DICL: Dynamic In-Context Learning for Automated Brick Schema Classification

22. FinAcumen: Financial Multimodal Reasoning via Self-Evolving Experience Memory Harness

23. Beyond Domains: Reusing Web Skills via Transferable Interaction Patterns

24. From Brewing to Resolution: Tracing the Internal Lifecycle of Code Reasoning in LLMs

25. Using Cognitive Models to Improve Language Model Simulation of Human Persuasion Games

26. FllumaOne: A Code-Native Multimodal CAD Dataset with Executable Programs and Kernel-Validated Feature Histories

27. EComAgentBench: Benchmarking Shopping Agents on Long-Horizon Tasks with Distributed Hidden Intent

28. LongWebBench: Evaluating Structural and Functional Webpage Generation in Long-Horizon Settings

29. Shattering the Autoregressive Curse: Dynamic Epistemic Entropy Orchestrated Erasable Reinforcement Learning for LLMs

30. DecoSearch: Complexity-Aware Routing and Plan-Level Repair for Text-to-SQL

31. WallZero: Mastering the Game of WallGo with Strategic Analysis

32. A homotopy-type-theoretic generalization of neurosymbolic inference

33. FlowRAG: Synergizing Explicit Reasoning via Frequency-Aware Multi-Granularity Graph Flow

34. StepGuard: Guarding Web Navigation via Single-Step Calibration

35. Structural Preservation and the Logical Expressiveness of Graph Neural Networks

36. MathVis-Fine: Aligning Visual Supervision with Necessity via Progressive Dependency-Guided Training for Multimodal Mathematical Reasoning

37. Learn to Quantify Social Interaction with Constraints for Pedestrian Walking

38. DiagFlowBench: Evaluating How Language Models Handle Off-Procedure Inputs in Grounded Diagnostic Dialogue

39. PreAct: Computer-Using Agents that Get Faster on Repeated Tasks

40. How Inference Compute Shapes Frontier LLM Evaluation

41. Small Initialization Matters for Large Language Models

42. MoCo-AIS: A Contrastive Learning Framework for Similarity Computation of Vessel Trajectories

43. STAR: SpatioTemporal Adaptive Reward Allocation for Text-to-Image RL Post-Training

44. LLM Consumer Behavior Theory: Foundations of a Novel Research Field

45. LegalHalluLens: Typed Hallucination Auditing and Calibrated Multi-Agent Debate for Trustworthy Legal AI

46. ProvenanceGuard: Source-Aware Factuality Verification for MCP-Based LLM Agents

47. PseudoBench: Measuring How Agentic Auto-Research Fuels Pseudoscience

48. Agentic AI-based Framework for Mitigating Premature Diagnostic Handoff and Silent Hallucination in Healthcare Applications

49. A Unified Framework for Context-Aware and Relation-Aware Graph Retrieval-Augmented Generation

50. IsabeLLM: Automated Theorem Proving Applied to Formally Verifying Consensus

51. Trust the Right Teacher: Quality-Aware Self-Distillation for GUI Grounding

52. First Proof Second Batch

53. Knowledge Reutilization in Meta-Reinforcement Learning

54. Your AI Travel Agent Would Book You a Bullfight: An Agentic Benchmark for Implicit Animal Welfare in Frontier AI Models

55. Memory as a Wasting Asset: Pricing Flash Endurance for Embodied Agents, and the Limits of Doing So

56. WEQA: Wearable hEalth Question Answering with Query-Adaptive Agentic Reasoning

57. Learning Cardiac Electrophysiology Digital Twins Through Agentic Discovery of Hybrid Structure

58. DRFLOW: A Deep Research Benchmark for Personalized Workflow Prediction

59. The Stanford EDGAR Filings Dataset: Reconstructing U.S. Corporate and Financial Disclosures into Layout-Faithful and Token-Efficient Pretraining Data

60. Fixed-Point Reasoners: Stable and Adaptive Deep Looped Transformers

61. EvolveNav: Proactive Preflection and Self-Evolving Memory for Zero-Shot Object Goal Navigation

62. Correct When Paired, Wrong When Split: Decoupling and Editing Modality-Specific Neurons in MLLMs

63. Towards Distributed Inference of LLMs on a P2P Network

64. PIVOT: Bridging Black-Scholes Implied-Volatility and Price Objectives via Differentiable J\"ackel Operator

65. KFTD: Koopman-Fourier Time-Differentiable Network for Continuous Ocean Spatiotemporal Forecasting

66. Extracting Semantics: LLM-Guided Automatic Population of Robot Ontology from URDF

67. Surveying GenAI-based Automation in Printed Circuit Board Design and Test

68. CMIP-Forge: An Agentic System that Retrieves, Computes, and Self-Reviews Climate Science

69. Comprehensive pKa Data Augmentation from Limited Real Data through an Engineered Models-Quantum Framework

70. HRDX: A Large-Scale Vector HD-Map Dataset

71. The Price of Anarchy in Disaggregated Inference

72. ParkingTransformer: LLM-Enhanced End-to-End Trajectory Planning for Autonomous Parking

73. ZIVARI-TLBO: A Zero-Cost Inter-Group Evaluated-Elite Relay Mechanism for Teaching-Learning-Based Optimization

74. ANEForge: Python for direct computation on the Apple Neural Engine

75. Software Delegation Contracts: Measuring Reviewability in AI Coding-Agent Work

76. Quantum Cinema: An Interactive Cinematic Exploration of Quantum Computing Hardware via Generative World Models

77. Prefill/Decode-Aware Evaluation of LLM Inference on Emerging AI Accelerators

78. Models Take Notes at Prefill: KV Cache Can Be Editable and Composable

79. Timestamp-Aware Spatio-Temporal Graph Contrastive Learning for Network Intrusion Detection