dag

GraMRAG is a new multi-agent RAG framework that uses graph memory and reinforcement learning to improve reasoning depth and memory structure in complex multimodal tasks, achieving state-of-the-art performance.

This paper introduces PATH, a framework for active curriculum refinement in reinforcement learning that leverages curriculum graph structure to improve robustness and generalization across diverse environments.

A Google engineer's talk highlights how AI agents are transforming software engineering, making traditional IDEs obsolete, with 85% of PRs at Google already shipped by AI agents and introducing concepts like DAGs for self-improving autonomous systems.

A technical blog post that walks through building a production-grade agentic harness around a basic LLM loop, covering typed tools, plan DAGs, tiered memory, verification hierarchies, budgets, and tracing.

This paper introduces GoT-CD, a Graph-of-Thoughts method for causal discovery that generates valid DAGs and shows that while structurally competitive with LLM baselines, structural fidelity alone does not guarantee fairness-faithful path-specific audits, highlighting the fragility of post-hoc path-specific fairness auditing.

A paper presents a structural approach to session compaction, modeling chat history as a DAG with snapshots and branches to trim bloat while preserving meaning, achieving a mean 20% token cut in coding sessions.

This paper presents a workload-aware cache eviction policy for multi-agent systems that uses recomputation cost, DAG dependency count, and agent invocation frequency to retain valuable cached entries, reducing latency by up to 64.7% over uncached baselines and 31.1% over the next best finite-capacity method.

BatchDAG introduces a system where an LLM generates typed directed acyclic graphs of operations for scalable ad-hoc analysis over enterprise data, achieving up to 47x reduction in LLM calls and sub-60-second query times over 50,000+ meetings.

The paper introduces a knowledge-centric framework for generating ComfyUI workflows by distilling hierarchical knowledge (pseudo-codes, skeletons, strategies) from real workflows and using LLMs to perform reasoning from task descriptions to executable structures, achieving higher node diversity and execution success rates.

DataFlow-Harness introduces a platform that guides LLM agents to construct editable DAG-based data pipelines through incremental mutations, achieving high pass rates and reduced costs compared to script-generation baselines.

A thread introducing the Atomic Task Graph (ATG), a DAG-based execution substrate that allows a small 8B model to outperform GPT-4 on complex agent tasks by storing plan structure in a graph instead of linear text, reducing context bloat and enabling localized failure recovery.

The Marble Curriculum, a comprehensive open-source dataset covering 1,590 primary school concepts with 3,221 connections across 8 subjects, has been released to enable building learning paths and AI-driven educational tools.

Proposes a finite-sample method for recovering the sparsest DAG in linear non-Gaussian acyclic models with latent confounders using higher-order cumulants, without restricting the number of latents.

A developer built a deterministic firewall in Rust for AI agents that runs under 5ms, using plans, MCP lists, and taint tracking to enforce security without hallucination or latency issues.

A blog post explaining PyTorch FX graphs, which are an intermediate representation used throughout the PyTorch 2.0 compile ecosystem. It covers the core objects Graph, Node, and GraphModule, and how to understand and work with them.

Introduces ReasoningFlow, a framework to capture discourse structures of large language model reasoning traces as directed acyclic graphs, enabling fine-grained analysis of reasoning behaviors like self-reflection and backtracking. Based on manual and automatic annotation of thousands of traces, it reveals structural similarities across models and that most erroneous steps do not contribute to final answers.

Yacine conducted a 1.5-hour in-depth interview with the founders of Paradigma, discussing how to use DAG (Directed Acyclic Graph) as the underlying infrastructure for autonomous research, covering core topics such as Agent operation, building large-scale public DAGs, and avoiding bad DAGs.

Interview discussing infrastructure for auto-research using DAGs, including how agents can execute DAGs and how to build large public DAGs.

PACER is a new scalable framework for causal discovery from large-scale interventional data that guarantees acyclicity by design, achieving up to two orders of magnitude speedups over penalty-based methods on benchmarks with thousands of variables.

GraphBit is a graph-based agentic framework that uses deterministic DAG orchestration with a Rust engine to eliminate hallucinations and infinite loops. It achieves 67.6% accuracy on GAIA benchmarks with zero framework-induced errors and low latency.