ICLR 2026 Recursive Self-Improvement

reddit_analysis_results.tsv

reddit_analysis_results.tsv

TamperBench_A_Systematic_Framework_to_Stress-Test_LLM_Safety_Under_Fine-Tuning_and_Tampering__smLtz7WID0.pdf

Depth_vs_Recursion_Outperforming_Transformers_in_Jigsaw_Reconstruction__1zDG1o16xB.pdf

Test-Time_Adaptation_via_Many-Shot_Prompting_Benefits_Limits_and_Pitfalls__Ne0s3UOzdB.pdf

Reference-Guided_Machine_Unlearning__0IbjWHRGMC.pdf

Cross-Family_Speculative_Prefill_Training-Free_Long-Context_Compression_with_Small_Draft_Models__7dNjMllnWe.pdf

Federation_over_Text__gh7aCT73cq.pdf

Build_Judge_Optimize_A_Blueprint_for_Continuous_Improvement_of_Multi-Agent_Consumer_Assistants__FySoHBWmt9.pdf

Real-Time_Procedural_Learning_From_Experience_for_AI_Agents__MKG4BaSieN.pdf

Orthogonal_Gradient_Projection_for_Continual_LLM_Unlearning__lb6Ce20kl5.pdf

Test-Time_Meta-Adaptation_with_Self-Synthesis__G0GE1xbR0w.pdf

Compute_as_Teacher_Turning_Inference_Compute_Into_Reference-Free_Supervision__WA6q2pNQhj.pdf

Constructive_Distortion_Improving_MLLMs_with_Attention-Guided_Image_Warping__sSwXGq1RRI.pdf

VerlTool_Towards_Holistic_Agentic_Reinforcement_Learning_with_Tool_Use__l5QPYxMEaH.pdf

Contrastive_Self-Refinement_for_Low-Cost_Adaptation_in_Real-World_Text-to-SQL__tgnbgt3Ctr.pdf

AutoHarness_Improving_LLM_Agents_by_Automatically_Synthesizing_a_Code_Harness__g9rEYVNn5T.pdf

Dynamic_Noise_Preference_Optimization__nT1gTSAf07.pdf

Aligned_but_Stereotypical_Understanding_and_Mitigating_Social_Bias_in_LLM-Driven_Text-to-Image_Models__RvgoRhan6d.pdf

Evaluating_AGENTS.md_Are_Repository-Level_Context_Files_Helpful_for_Coding_Agents__0DyJeJ3iia.pdf

ESDAE_Evaluating_Synthetic_Data_for_Agent_Evaluation__DlhLLVhrZB.pdf

Beyond_Solving_A_Closer_Look_at_LLMs_as_Solution_Verifiers__Hv5hDfbhuB.pdf

Actor-Curator_Scalable_Policy-driven_Curriculum_Learning_for_RL_Post-Training__JYJXWnujvR.pdf

Learning_to_Evolve_Scaling_Open-Ended_Discovery_with_Relative-Progress_RL__WnZHbe1Gu0.pdf

Inference-Time_Scaling_in_Diffusion_Models_through_Iterative_Partial_Refinement__QopjICzGwr.pdf

Self-Improving_Clinical_Reasoning_via_Textual_Gradients__VOa0lDJ5Ui.pdf

Agentic_Context_Engineering_Evolving_Contexts_for_Self-Improving_Language_Models__f1eKYFvMNu.pdf

Learning_What_to_Learn_Curriculum_Curation_for_Test-Time_Agent_Learning__TRQLuxgxBN.pdf

Refining_Large_Language_Models_with_Self-Generated_Data_Through_Iterative_Training__81P20Qg44P.pdf

AlphaApollo_A_System_for_Deep_Agentic_Reasoning__UcIFJPXqrB.pdf

Teaching_Models_to_Teach_Themselves_Reasoning_at_the_Edge_of_Learnability__HuLoDnvC29.pdf

TextBO_Bayesian_Optimization_in_Language_Space_for_Eval-Efficient_Self-Improving_AI__h0XYERyeL6.pdf

SAGE_Self-play_Adversarial_Games_Enhance_Large_Language_Model_Reasoning_Capabilities__b3dPMokQki.pdf

Self-Improvement_via_Fast_Tree-search__wZMNXHPYcO.pdf

Federated_Agent_Reinforcement_Learning__h4iPWHGapI.pdf

Reward_Hacking_in_Self-Improving_Code_Agents__ikrQWGgxYg.pdf

Adaptive_Decoding_via_Test-Time_Policy_Learning_for_Self-Improving_Generation__7fIqKZ92Fy.pdf

Leveraging_Suboptimal_and_Noisy_Trajectories_for_Goal-Conditional_Offline_RL__jfi88XbVsn.pdf

Self-CriTeach_LLM_Self-Teaching_and_Self-Critiquing_for_Improving_Robotic_Planning_via_Automated_Domain_Generation__8I0n20ufAy.pdf

Just_Enough_Learning_GRPO-Guided_Controllers_for_Hyperparameter_Sweeps__kKWSQsYgpa.pdf

Discover_the_distinguishing_and_effective_reasoning_patterns_among_LLMs_via_an_LLM__z9MOHU8RjL.pdf

Generative_Recursive_Reasoning_Models__Vxu6kcIjwV.pdf

Improved_Iterative_Refinement_for_Chart-to-Code_Generation_via_Structured_Instruction__KgkMSbYoL1.pdf

Duel-Evolve_Pairwise_Preference_Black-Box_Optimization_of_LLM_Responses__38t39AFZPE.pdf

Soft_Mellowmax_Monte_Carlo_Planning__SeZCSEGZTq.pdf

Language-Guided_Expertise_Evolution_for_Protein_Optimization__NeNWeyfUJj.pdf

Vision-Guided_Iterative_Refinement_for_Frontend_Code_Generation__gF6jmAsbpZ.pdf

A_Framework_for_Prompt_Optimization_and_Translation_Across_Foundation_Models__gOTCn5uZeE.pdf

MimicAgent_Learning_Quadruped_Skills_via_Text-to-Trajectory_Generation__1iAEFtFQ9M.pdf

Simple_Baselines_are_Competitive_with_Code_Evolution__QSWFqDcveB.pdf

TangramSR_A_Benchmark_for_Recursive_Self-Improvement_In_Continuous_Geometric_Reasoning__5XGfbXlFmA.pdf

CircuitBuilder_From_Polynomials_to_Circuits_via_Reinforcement_Learning__JNsTWIukjQ.pdf

Feedback_Descent_Open-Ended_Text_Optimization_via_Pairwise_Comparison__A0dlCcZ4hT.pdf

Do_Depth-Grown_Models_Overcome_the_Curse_of_Depth__Enqas0eZF6.pdf

Agent0-VL_Exploring_Self-Evolving_Agent_for_Tool-Integrated_Vision-Language_Reasoning__JKx0zkUBsQ.pdf

OMEGA_Optimizing_Machine_learning_by_Evaluating_Generated_Algorithms__4TUzVEzVdu.pdf

Reasoning_as_Gradient_Scaling_MLE_Agents_Beyond_Tree_Search__TnjlvLY30w.pdf

Intelligent_Robot_Manipulation_Requires_Self-Directed_Learning__hCjgh8jxKl.pdf

Rethinking_Machine_Unlearning_Models_Designed_to_Forget_via_Key_Deletion__gGH3Xp1lHR.pdf

Emergent_temporal_abstractions_in_autoregressive_models_enable_hierarchical_reinforcement_learning__5UbKomO77O.pdf

Log-Augmented_Generation_Scaling_Test-Time_Reasoning_with_Reusable_Computation__32xd2HdB9m.pdf

Differentiable_Evolutionary_Reinforcement_Learning__uXTXcZ615k.pdf

Self-Improving_VLM_Judges_Without_Human_Annotations__8hYSvUpJBA.pdf

In-Context_Adaptation__f58uDOwLaq.pdf

LLM-FE_Automated_Feature_Engineering_for_Tabular_Data_with_LLMs_as_Evolutionary_Optimizers__KbGkWv7VEk.pdf

Your_Self-Play_Algorithm_is_Secretly_an_Adversarial_Imitator__rL0GEyoMvE.pdf

Self-Adapting_Agents_for_Automating_Research_Coding_Workflows__ihcRmUkXHF.pdf

Self-EvolveRec_Self-Evolving_Recommender_Systems_with_LLM-based_Directional_Feedback__KCcpSuocz0.pdf

Shape_of_Thought_When_Distribution_Matters_More_than_Correctness_in_Reasoning_Tasks__Oc5R3D2TfE.pdf

Verifying_the_Verifiers_Failure_Attribution_for_Agentic_Benchmark_Diagnostics_and_Training_Data_Curation__iRhaK8PsuB.pdf

Correct_Reasoning_Paths_Visit_Shared_Decision_Pivots__OlMSrldTNe.pdf

CoT-Seg_Rethinking_Segmentation_with_Chain-of-Thought_Reasoning_and_Self-Correction__BtHRvwzJj8.pdf

POLARIS_A_Godel_Agent_Framework_for_Small_Language_Models_Through_Experience_Abstracted_Policy_Repair__wlwkLVvB0W.pdf

MAPPA_Scaling_Multiagent_Systems_with_Process_Rewards__s06wgoO65a.pdf

Unlocking_Intrinsic_Self-Reflection_for_LLM_Preference_Policy_Optimization__0ZGL40jRKE.pdf

A_Task-Centric_Theory_for_Iterative_Self-Improvement_with_Easy-to-Hard_Curricula__EEc4Pn2GSa.pdf

Reasoning_Within_the_Mind_Dynamic_Multimodal_Interleaving_in_Latent_Space__fWrhAOQZ5A.pdf

Theory-Driven_Modeling_and_LLM-Guided_Evolution_for_Power_System_Scheduling__djdSNvWpJ7.pdf

SkillRL_Evolving_Agents_via_Recursive_Skill-Augmented_Reinforcement_Learning__56D2hjARkn.pdf

One-Step_Video_Depth_Estimation_via_Self-Distillation__ucRjPd9HVU.pdf

Reasoning_Cache_Learning_to_Extrapolate_to_Long_Lengths_via_Short-Length_RL__DROMQyqM52.pdf

Residual_Off-Policy_RL_for_Finetuning_Behavior_Cloning_Policies__9m4kM3bK2b.pdf

Structure_Enables_Effective_Self-Localization_of_Errors_in_LLMs__QbL99Fyqsl.pdf

RFTF_Reinforcement_Fine-tuning_for_Vision-language-action_Models_with_Temporal_Feedback__mBgcG43j7a.pdf

Escaping_Model_Collapse_via_Synthetic_Data_Verification__YzDC5hjGUM.pdf

SAHOO_Safeguarded_Alignment_for_High-Order_Optimization_Objectives_in_Recursive_Self-Improvement__OAFPpQO0H9.pdf

Unrolled_Policy_Iteration_for_Tiny_Recursive_Models__rIzREKws05.pdf

Agent0_Unleashing_Self-Evolving_Agents_from_Zero_Data_via_Tool-Integrated_Reasoning__hYYeOl58xi.pdf

Learning_to_Continually_Learn_via_Meta-learning_Agentic_Memory_Designs__sOq52KnJmR.pdf

PostTrainBench_Can_LLM_Agents_Automate_LLM_Post-Training__FJKOIxkUxo.pdf

Contextual_Drag_How_Errors_in_the_Context_Affect_LLM_Reasoning__zpiYsPVDlV.pdf

Can_Current_Language_Models_Close_the_Discovery_to_Application_Loop__BAdK20xfqj.pdf

Interestingness_as_an_Inductive_Heuristic_for_Future_Compression_Progress__6GTlSlWW9C.pdf

Adaptive_Meta-Curriculum_for_Test-Time_Self-Improvement__GjoUJTfXiW.pdf

SimpleMem_Efficient_Lifelong_Memory_for_LLM_Agents__oYHelQ3Edd.pdf

Self-Improving_World_Models_via_Asymmetric_Forward-Inverse_Consistency__ajcjip0yFR.pdf

Towards_Execution-Grounded_Automated_AI_Research__gpLJamvbsK.pdf

Lang-PINN_From_Language_to_Physics-Informed_Neural_Networks_via_a_Multi-Agent_Framework__q5qN3oQ4D1.pdf

VLAW_Iterative_Co-Improvement_of_Vision-Language-Action_Policy_and_World_Model__Ro0eQ0ly3q.pdf

Presenting_a_Paper_is_an_Art_Self-Improvement_Aesthetic_Agents_for_Academic_Presentations__d40v7Qcpi4.pdf

GASP_Guided_Asymmetric_Self-Play_For_Coding_LLMs__NYrOkAfDkP.pdf

Self-Evolving_Rubrics_Interpretable_Instance-Level_Criteria_for_Scalable_RL__aA2PXFH2Cp.pdf

Language_Self-Play_For_Data-Free_Training__uB8YQHNsh6.pdf

Self-Improving_Vision-Language-Action_Models_with_Data_Generation_via_Residual_RL__shqo2VPfc7.pdf

ACE_Self-Evolving_LLM_Coding_Framework_Adversarial_Unit_Test_Generation_and_Preference_Optimization__ecKAmz5vlO.pdf

CausalEvolve_Towards_Open-Ended_Discovery_with_Causal_Scratchpad__14jSctSh0D.pdf

Knowledge_is_Not_Enough_Injecting_RL_Skills_for_Continual_Adaptation__D93migx9av.pdf

Test-Time_Self-Distillation__iEmRSwdzyw.pdf

Anchored_Self-Play_for_Code_Repair__lTbBFAoPSA.pdf

From_Growing_to_Looping_A_Unified_View_of_Iterative_Computation_in_LLMs__yIDgVx3OoN.pdf

Tiny_Autoregressive_Recursive_Models__aY5kmaNrwB.pdf

Can_Language_Models_Discover_Scaling_Laws__Nj6VGY4dej.pdf