Learning GenAI via SOTA Papers

Title: Structural Enforcement of Goal Integrity in AI Agents via Separation-of-Powers Architecture Source: http://arxiv.org/abs/2604.23646v1 Summary: This research introduces the Policy-Execution-Authorization (PEA) architecture, a foundational system-level primitive that decouples agent intent from execution to structurally prevent agentic misalignment. By moving beyond probabilistic model-level...

Title: DLM: Unified Decision Language Models for Offline Multi-Agent Sequential Decision Making Source: http://arxiv.org/abs/2604.23557v1 Summary: This paper proposes a unified framework that treats multi-agent decision-making as a dialogue-style sequence prediction problem, enabling robust zero-shot generalization across heterogeneous environments. It establishes a foundational approach for scali...

Title: From Coarse to Fine: Self-Adaptive Hierarchical Planning for LLM Agents Source: http://arxiv.org/abs/2604.23194v1 Summary: This paper introduces AdaPlan-H, a novel agentic reasoning framework that enables LLM agents to dynamically adjust planning granularity based on task complexity. It provides a foundational primitive for long-horizon task execution by mimicking human progressive refineme...

Title: ArgRE: Formal Argumentation for Conflict Resolution in Multi-Agent Requirements Negotiation Source: http://arxiv.org/abs/2604.23124v1 Summary: ArgRE establishes a foundational framework for multi-agent coordination by embedding formal Dung-style abstract argumentation into the agentic reasoning loop for conflict resolution. It moves multi-agent negotiation from heuristic synthesis to a prin...

Title: Tool Attention Is All You Need: Dynamic Tool Gating and Lazy Schema Loading for Eliminating the MCP/Tools Tax in Scalable Agentic Workflows Source: http://arxiv.org/abs/2604.21816v1 Summary: This paper introduces a middleware-layer mechanism that reduces tool-related token overhead by 95% through dynamic gating and lazy schema loading. It provides a critical efficiency breakthrough for scal...

Title: DryRUN: On the Role of Public Tests in LLM-Driven Code Generation Source: http://arxiv.org/abs/2604.21598v1 Summary: DryRUN introduces a novel agentic reasoning loop where LLMs autonomously generate inputs and simulate execution traces to self-correct code without relying on human-provided test cases. This removes a primary bottleneck for autonomous software engineering and mitigates the &#...

Title: Image Generators are Generalist Vision Learners Source: http://arxiv.org/abs/2604.20329v1 Summary: This paper demonstrates that image generation pretraining serves as a unified foundation for both visual creation and zero-shot understanding, rivaling domain-specific specialists across diverse 2D and 3D tasks. It proposes a paradigm shift where generative models act as generalist vision lear...

Title: Automatic Ontology Construction Using LLMs as an External Layer of Memory, Verification, and Planning for Hybrid Intelligent Systems Source: http://arxiv.org/abs/2604.20795v1 Summary: The paper introduces a hybrid architecture that integrates LLMs with an external ontological memory layer to provide semantically grounded reasoning and long-term persistence. This framework enables the automa...

Title: Taming Actor-Observer Asymmetry in Agents via Dialectical Alignment Source: http://arxiv.org/abs/2604.19548v1 Summary: This paper introduces ReTAS, a dialectical reasoning framework that mitigates Actor-Observer Asymmetry in multi-agent systems by synthesizing conflicting actor and observer perspectives into objective consensus. By enforcing perspective-invariant reasoning through dialectic...

Title: OLLM: Options-based Large Language Models Source: http://arxiv.org/abs/2604.19087v1 Summary: OLLM introduces a novel architectural modification that replaces standard next-token prediction with a set of learned options indexed by a discrete latent variable, significantly enhancing diversity and controllability. This shift to latent-space policy selection provides a foundational primitive fo...

Title: Sessa: Selective State Space Attention Source: http://arxiv.org/abs/2604.18580v2 Summary: This paper introduces a novel architectural primitive that integrates attention into a recurrent feedback path, achieving power-law memory tails for superior long-context information preservation. It represents a significant breakthrough by combining the strengths of Transformers and State-Space Models...

Title: Agent-World: Scaling Real-World Environment Synthesis for Evolving General Agent Intelligence Source: http://arxiv.org/abs/2604.18292v1 Summary: Agent-World provides a foundational framework for scaling general agent intelligence by autonomously synthesizing thousands of real-world environment themes and tasks for continuous self-evolution. The co-evolution of agent policies and environment...

Title: Hive: A Multi-Agent Infrastructure for Algorithm- and Task-Level Scaling Source: http://arxiv.org/abs/2604.17353v1 Summary: Hive provides a foundational infrastructure for scaling agentic systems at the algorithm and task levels, introducing key mechanisms like Logits Cache and Agent-Aware Scheduling to optimize multi-agent resource allocation. It represents a significant breakthrough in ef...

Title: Compiling Deterministic Structure into SLM Harnesses Source: http://arxiv.org/abs/2604.17450v1 Summary: This paper introduces Semantic Gradient Descent (SGDe), a novel PAC-learning-based framework that compiles complex agentic workflows into optimized, deterministic execution plans for small language models. It establishes a new architectural primitive for agent design by leveraging frontie...

Title: The Topological Trouble With Transformers Source: http://arxiv.org/abs/2604.17121v1 Summary: This research identifies a fundamental topological limit in the Transformer architecture's ability to track dynamic states due to its feedforward nature, which exhausts model depth as states evolve. It advocates for a shift towards recurrent or continuous-thought architectures as essential primi...

Title: The Consensus Trap: Rescuing Multi-Agent LLMs from Adversarial Majorities via Token-Level Collaboration Source: http://arxiv.org/abs/2604.17139v1 Summary: This paper introduces Token-Level Round-Robin Collaboration, a novel multi-agent interaction paradigm that interleaves token generation to overcome the brittle aggregation of majority voting. It provides a mathematical proof that this non...

Title: Neurosymbolic Repo-level Code Localization Source: http://arxiv.org/abs/2604.16021v1 Summary: This work presents LogicLoc, a neurosymbolic agentic framework that integrates LLMs with Datalog for deterministic structural reasoning in codebase analysis. It represents a foundational shift toward verifiable agentic workflows by offloading complex structural traversals to a symbolic engine, dras...

Title: Structured Abductive-Deductive-Inductive Reasoning for LLMs via Algebraic Invariants Source: http://arxiv.org/abs/2604.15727v1 Summary: This paper introduces a symbolic reasoning scaffold that operationalizes tripartite inference (abduction, deduction, induction) to enforce logical consistency in LLMs. By using algebraic invariants like the "Weakest Link" bound, it provides a foun...

Title: SGA-MCTS: Decoupling Planning from Execution via Training-Free Atomic Experience Retrieval Source: http://arxiv.org/abs/2604.14712v1 Summary: This work presents a novel framework that amortizes the high cost of inference-time search by casting LLM planning as non-parametric retrieval of symbolic 'SGA atoms.' By enabling System 2 reasoning depth at System 1 speeds without task-specif...

Title: Autogenesis: A Self-Evolving Agent Protocol Source: http://arxiv.org/abs/2604.15034v1 Summary: This paper introduces the Autogenesis Protocol (AGP), a foundational framework for self-evolving multi-agent systems that standardizes the lifecycle and evolution of agentic resources. It provides a structured approach to decoupling agent evolution from execution, enabling scalable and auditable i...

Title: AIBuildAI: An AI Agent for Automatically Building AI Models Source: http://arxiv.org/abs/2604.14455v1 Summary: This paper introduces a hierarchical multi-agent framework that automates the full lifecycle of AI model development, achieving human-level performance on the MLE-Bench. It represents a foundational shift towards autonomous, self-improving agentic systems capable of complex enginee...

Title: Bridging MARL to SARL: An Order-Independent Multi-Agent Transformer via Latent Consensus Source: http://arxiv.org/abs/2604.13472v1 Summary: The provided text introduces the Consensus Multi-Agent Transformer (CMAT), a novel framework designed to improve cooperative multi-agent reinforcement learning (MARL) by reformulating it as a hierarchical single-agent problem. While traditional models o...

Paper Link: https://arxiv.org/abs/2604.12632 Summary: The provided sources introduce Calibration-Aware Policy Optimization (CAPO), a new reinforcement learning framework designed to improve both the accuracy and reliability of Large Language Models (LLMs). The research identifies a critical flaw in existing methods like Group Relative Policy Optimization (GRPO), which frequently causes models to b...

Paper Link: https://arxiv.org/abs/2604.04921 Summary: The provided sources introduce TriAttention, a novel KV cache compression technique designed to enhance the efficiency of Large Language Models during long-context reasoning. By identifying that query and key vectors concentrate around stable centers in the pre-RoPE space, the researchers developed a trigonometric series to predict and retain t...

Paper Link: Summary: The provided sources introduce SOAR (Self-Optimization via Asymmetric RL), a meta-reinforcement learning framework designed to help large language models overcome reasoning plateaus. While standard training methods often stall on problems the model cannot already solve, this system uses an asymmetric teacher-student setup where the teacher generates synthetic "stepping st...