PRISM-0: A Predicate-Rich Scene Graph Generation Framework for Zero-Shot Open-Vocabulary Tasks

The introduction of WISE-Flow, a workflow-centric framework, aims to enhance the capabilities of large language model (LLM)-based conversational agents by converting historical service interactions into reusable procedural experiences. This approach addresses the common issues of error-proneness and variability in agent performance across different tasks.

A recent study has investigated the dynamics of Large Language Model (LLM) agent reviewers within an Elo-ranked review system, utilizing real-world conference paper submissions. The research involved multiple LLM reviewers with distinct personas engaging in multi-round review interactions, moderated by an Area Chair, and highlighted the impact of Elo ratings and reviewer memory on decision-making accuracy.

Salience-SGG introduces a novel framework for Scene Graph Generation (SGG) that addresses the bias in traditional models caused by a long-tailed distribution of predicate classes. By utilizing an Iterative Salience Decoder (ISD) and semantic-agnostic salience labels, it enhances spatial understanding and improves performance on datasets like Visual Genome, Open Images V6, and GQA-200.

A new framework for matrix deflation has been proposed, utilizing an agentic approach where a Large Language Model (LLM) generates rank-1 Singular Value Decomposition (SVD) updates, while a Vision Language Model (VLM) evaluates these updates, enhancing solver stability through in-context learning and strategic permutations. This method was tested on various matrices, demonstrating promising results in noise reduction and accuracy.