AI Papers of the Week
Every paper worth reading in AI, hand-picked one week at a time.
AgentArk
AgentArk distills multi-agent debate dynamics into a single LLM, transferring the reasoning and self-correction abilities of multi-agent systems into one model at training time. Three hierarchical distillation strategies (reasoning-enhanced SFT, trajectory-based augmentation, and process-aware distillation with a process reward model) yield an average 4.8% improvement over single-agent baselines across math and reasoning benchmarks, approaching full multi-agent performance at a fraction of the inference cost. Cross-family distillation (e.g., Qwen3-32B to LLaMA-3-8B) produces the largest gains, suggesting heterogeneous architectures benefit most from transferred reasoning signals.
Semi-Autonomous Mathematics Discovery with Gemini
This paper from Google DeepMind presents a case study in semi-autonomous mathematics discovery using Aletheia, a specialized math research agent built on Gemini Deep Think. The team systematically evaluated 700 open conjectures from Bloom’s Erdos Problems database, combining AI-driven natural language verification with human expert evaluation, and addressed 13 previously open problems.
A-RAG
A-RAG is an agentic RAG framework that gives LLMs direct access to hierarchical retrieval interfaces instead of relying on fixed retrieval algorithms or predefined workflows. The agent autonomously decides what to search, at which granularity, and when to stop - representing a paradigm shift from static retrieval pipelines to truly agentic information gathering.
Accelerating Scientific Research with Gemini
A collection of case studies from Google Research showing how researchers used Gemini Deep Think to solve open problems, refute conjectures, and generate new proofs across theoretical computer science, information theory, cryptography, optimization, economics, and physics. The paper extracts a practical playbook of recurring techniques, including iterative refinement, cross-disciplinary knowledge transfer, counterexample search, and neuro-symbolic verification loops where the model autonomously writes and executes code to validate derivations. Notable results include identifying a fatal flaw in a cryptography preprint on SNARGs, resolving the Courtade-Kumar conjecture in information theory, and proving that the simplex is optimal for Euclidean Steiner trees.
Kimi K2.5: Visual Agentic Intelligence
Kimi K2.5 is an open-source multimodal agentic model from Moonshot AI that jointly optimizes text and vision capabilities through native multimodal pretraining on 15 trillion mixed tokens, zero-vision SFT, and joint reinforcement learning. K2.5 also introduces Agent Swarm, a parallel agent orchestration framework that dynamically decomposes complex tasks into concurrent subtasks, reducing latency by up to 4.5x over single-agent baselines.
Reinforcement Learning via Self-Distillation
This paper introduces Self-Distillation Policy Optimization (SDPO), an on-policy RL algorithm that converts rich textual feedback from verifiable environments into dense credit assignment without requiring an external teacher model. SDPO uses the current model conditioned on feedback as a “self-teacher” to retrospectively identify mistakes in its own rollouts, substantially outperforming GRPO across scientific reasoning, tool use, and competitive programming.
Team of Rivals: Orchestrating Reliable AI Agents
This paper proposes organizing AI agents into corporate-style teams with strict role boundaries and opposing incentives (planners, executors, critics, experts) to achieve reliability through careful orchestration of imperfect components. A remote code executor separates reasoning from data transformations, preventing raw tool outputs from contaminating agent context windows. The system achieves over 90% internal error interception before user exposure while maintaining acceptable latency tradeoffs.
TTT-Discover: Learning to Discover at Test Time
TTT-Discover introduces test-time training for scientific discovery, performing reinforcement learning at test time so the LLM can continue to train with experience specific to the test problem. Unlike prior work like AlphaEvolve that prompts a frozen LLM, this approach enables the model itself to improve while attempting to solve hard problems.
Reasoning Models Generate Societies of Thought
This paper reveals that enhanced reasoning in models like DeepSeek-R1 and QwQ-32B emerges not from extended computation alone, but from simulating multi-agent-like interactions - a “society of thought” - enabling diversification and debate among internal cognitive perspectives with distinct personality traits and domain expertise.
Rethinking Multi-Agent Workflows
This paper challenges the assumption that complex tasks require multiple specialized AI agents, demonstrating that a single LLM agent, through iterative dialogue, can match the performance of homogeneous multi-agent workflows while gaining efficiency from KV cache reuse.
Task-Decoupled Planning for Long-Horizon Agents
Task-Decoupled Planning (TDP) is a training-free framework that restructures agent planning by decomposing tasks into a directed acyclic graph of sub-goals using three components: Supervisor, Planner, and Executor. By isolating reasoning to individual subtasks through scoped contexts, TDP prevents error cascading and reduces token consumption by up to 82% while outperforming baselines on TravelPlanner, ScienceWorld, and HotpotQA.
Self-Evolving Search Agents Without Training Data
Dr. Zero introduces a framework for developing multi-turn search agents that improve themselves autonomously without labeled training data. A proposer generates diverse questions to train a solver initialized from the same base model, creating a self-evolution loop with automated curriculum difficulty scaling.
Active Context Compression for LLM Agents
Focus introduces an agent-centered architecture that enables LLM agents to autonomously manage their own memory by deciding when to consolidate learnings into a persistent “Knowledge” block and actively prune raw interaction history. The design is inspired by the biological navigation patterns of Physarum polycephalum (slime mold).
Agent-as-a-Judge
This comprehensive survey traces the evolution from LLM-based evaluation to agentic evaluation approaches, establishing the first taxonomy for this paradigm shift. As evaluation tasks grow more intricate and specialized, traditional single-pass language model judges become insufficient.
Ministral 3
Mistral AI releases Ministral 3, a family of compact language models (3B, 8B, 14B parameters) designed for compute and memory-constrained applications from mobile to edge deployments. Created through Cascade Distillation (iterative pruning with continued training), each size offers pretrained, instruction-finetuned, and reasoning variants with integrated image understanding, released under Apache 2.0.
MemRL
MemRL enables LLM agents to improve continuously without retraining by separating a frozen model’s reasoning from an evolving memory system. A Two-Phase Retrieval mechanism filters candidates by semantic relevance, then ranks them using learned Q-values that improve through trial-and-error, outperforming existing methods on HLE, BigCodeBench, ALFWorld, and Lifelong Agent Bench.
On the Slow Death of Scaling
This essay by Sara Hooker challenges the decade-long assumption that scaling compute always leads to better AI performance. It argues that the relationship between training compute and performance is highly uncertain and rapidly changing, with smaller models now routinely outperforming much larger ones.
Recursive Language Models
Recursive Language Models (RLMs) are a general inference strategy that allows LLMs to process arbitrarily long prompts by treating them as part of an external environment. Rather than feeding long contexts directly into the model, RLMs load the prompt as a variable in a Python REPL and let the LLM programmatically examine, decompose, and recursively call itself over snippets.
Nemotron-Cascade
Nemotron-Cascade introduces cascaded domain-wise reinforcement learning (Cascade RL) to build general-purpose reasoning models capable of operating in both instruct and deep thinking modes. Rather than blending heterogeneous prompts from different domains, Cascade RL orchestrates sequential, domain-wise RL stages that reduce engineering complexity while delivering state-of-the-art performance.
GDPO
GDPO addresses a critical flaw in training language models with multiple competing objectives. The authors discover that when applying Group Relative Policy Optimization (GRPO) to multi-reward settings, normalizing distinct rollout reward combinations causes them to collapse into identical advantage values, degrading training signal quality and stability.
Confucius Code Agent
Confucius Code Agent (CCA) is a software engineering agent designed to operate on large-scale codebases. Built on the Confucius SDK, it introduces a three-axis design philosophy separating Agent Experience (AX), User Experience (UX), and Developer Experience (DX) to enable robust multi-step reasoning and modular tool use.
Geometric Memory in Sequence Models
This paper identifies a dramatically different form of how deep sequence models store factual information called geometric memory, contrasting with the traditional associative memory view. Models synthesize embeddings encoding global relationships between all entities, even ones that never co-occur in training.
Universal Reasoning Model
This paper investigates why universal transformers excel at complex reasoning tasks like ARC-AGI. The key finding: performance gains come primarily from recurrent inductive bias and strong nonlinear components rather than elaborate architectural designs.
Spacing Effect for Generalization
Researchers from Tsinghua University investigate how the spacing effect - a well-documented learning principle where spaced intervals between training improve retention - can enhance generalization in both biological and artificial neural networks.