graph-based

A discussion highlighting the use of dynamic workflows and graph-based approaches with LLMs to manage context and avoid confirmation bias in agent interactions.

MileGPO proposes a method for credit assignment in long-horizon LLM agents using milestone inference with local evidence, achieving state-of-the-art performance on ALFWorld and WebShop benchmarks.

This paper proposes CONFER, a graph-based conflict-aware evidence negotiation framework for weakly supervised multimodal emotion recognition, addressing self-report unreliability and cross-modal conflict. It achieves competitive accuracy on AMIGOS, MAHNOB-HCI, and DEAP benchmarks.

Introduces Project2Task, a graph-guided project-level planning layer for autonomous research systems that decomposes a macro project brief into bounded, dependency-aware research tasks with explicit contribution ownership. Evaluations show improved portfolio quality and downstream task accuracy.

This paper introduces PatTree, a graph-based multimodal patient representation that automatically structures heterogeneous clinical data for medical classification tasks. It achieves state-of-the-art performance on the ADNI-1 cohort with 98.5% balanced accuracy for Alzheimer's disease classification.

A computer science paper presents a graph-based workflow serving engine that unifies agent operations into a global wGraph, using dynamic graph synthesis and differential KV-cache reuse to boost agent accuracy by 4.95% while cutting GPU memory usage by 4x.

Google released ADK 2.0, an open-source, code-first Python framework for building AI agents with graph-based orchestration and agent-to-agent delegation, aiming to replace expensive enterprise agent frameworks.

This paper proposes a graph-based framework that combines weak supervision with propagation graph analysis to detect and analyze disinformation narratives in Telegram ecosystems, focusing on Russian and Ukrainian channels.

Introduces TSRouter, a graph-based dynamic routing framework that selects the optimal modality (LLM or VLM) and model for time-series reasoning tasks, achieving 16-46% relative improvement over baselines and demonstrating zero-shot generalization.

This paper presents a training-free graph-based framework for reading order inference in complex document layouts, using language model signals and a max-regret inference rule. The method significantly outperforms existing baselines on historical manuscripts and multi-column benchmarks, achieving 95% edge accuracy on wrap-around Glossa layouts.

Google announces ADK for Go 2.0, introducing a graph-based workflow engine, human-in-the-loop primitives, and dynamic orchestration for building production-ready multi-agent applications.

This paper introduces MMG-Pop, a unified benchmark for multi-modal graph-based social media popularity prediction, and proposes MMG-PopNet, a model that jointly models multimodal content and temporal social interactions. Experiments on Bluesky and Reddit datasets demonstrate superior performance and provide insights into cross-platform generalization and multi-task prediction.

ScholarSum is a hierarchical reflective graph-based framework for scientific abstractive summarization that emulates a student–teacher writing process. It uses a hierarchical knowledge graph to capture global structure, generates an initial draft, and iteratively refines it via evidence retrieval and teacher-like review to improve both fluency and factual faithfulness.

The paper proposes GTBP, a graph-based back-propagation framework for context adaptation in multi-LLM agentic systems, which improves prompt optimization with theoretical convergence guarantees and outperforms existing methods on benchmarks.

TIGER is an inference-time framework that mitigates hallucinations in multimodal generation by extracting observation and claim graphs and assigning risk scores to repair unsupported facts. It reduces unsupported content across image-to-text, image+text-to-text, audio-to-text, and video-to-text tasks.

PropGuard is a propagation-aware framework for safeguarding LLM-based multi-agent systems (LLM-MAS) from malicious instructions that propagate across agents and rounds. It constructs a dual-view spatio-temporal graph and uses a GE-GRPO trained inspector to detect and remediate suspicious propagation subgraphs.

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.

Researchers from the University of British Columbia propose an unsupervised graph-based system for organizing arguments from online debates by constructing interaction graphs and applying community detection to reveal diverse viewpoint distributions. The approach requires no training data and aims to help users navigate complex argumentative landscapes and combat filter bubbles.