AgentSquare: Automatic LLM Agent Search in Modular Design Space

Recent advancements in Large Language Models (LLMs) have led to a rapid growth of agentic systems capable of handling a wide range of complex tasks. However, current research largely relies on manual, task-specific design, limiting their adaptability to novel tasks. In this paper, we introduce a new research problem: Modularized LLM Agent Search (MoLAS). We propose a modular design space that abstracts existing LLM agent designs into four fundamental modules with uniform IO interface: Planning, Reasoning, Tool Use, and Memory. Building on this design space, we present a novel LLM agent search framework called AgentSquare, which introduces two core mechanisms, i.e., module evolution and recombination, to efficiently search for optimized LLM agents. To further accelerate the process, we design a performance predictor that uses in-context surrogate models to skip unpromising agent designs. Extensive experiments across six benchmarks, covering the diverse scenarios of web, embodied, tool use and game applications, show that AgentSquare substantially outperforms hand-crafted agents, achieving an average performance gain of 17.2% against best-known human designs. Moreover, AgentSquare can generate interpretable design insights, enabling a deeper understanding of agentic architecture and its impact on task performance. We believe that the modular design space and AgentSquare search framework offer a platform for fully exploiting the potential of prior successful designs and consolidating the collective efforts of research community. Code repo is available at https://github.com/tsinghua-fib-lab/AgentSquare.

Fréchet Video Motion Distance: A Metric for Evaluating Motion Consistency in Videos

Significant advancements have been made in video generative models recently. Unlike image generation, video generation presents greater challenges, requiring not only generating high-quality frames but also ensuring temporal consistency across these frames. Despite the impressive progress, research on metrics for evaluating the quality of generated videos, especially concerning temporal and motion consistency, remains underexplored. To bridge this research gap, we propose Fr\'echet Video Motion Distance (FVMD) metric, which focuses on evaluating motion consistency in video generation. Specifically, we design explicit motion features based on key point tracking, and then measure the similarity between these features via the Fr\'echet distance. We conduct sensitivity analysis by injecting noise into real videos to verify the effectiveness of FVMD. Further, we carry out a large-scale human study, demonstrating that our metric effectively detects temporal noise and aligns better with human perceptions of generated video quality than existing metrics. Additionally, our motion features can consistently improve the performance of Video Quality Assessment (VQA) models, indicating that our approach is also applicable to unary video quality evaluation. Code is available at https://github.com/ljh0v0/FMD-frechet-motion-distance.

Forecasting Post-Wildfire Vegetation Recovery in California using a Convolutional Long Short-Term Memory Tensor Regression Network

The study of post-wildfire plant regrowth is essential for developing successful ecosystem recovery strategies. Prior research mainly examines key ecological and biogeographical factors influencing post-fire succession. This research proposes a novel approach for predicting and analyzing post-fire plant recovery. We develop a Convolutional Long Short-Term Memory Tensor Regression (ConvLSTMTR) network that predicts future Normalized Difference Vegetation Index (NDVI) based on short-term plant growth data after fire containment. The model is trained and tested on 104 major California wildfires occurring between 2013 and 2020, each with burn areas exceeding 3000 acres. The integration of ConvLSTM with tensor regression enables the calculation of an overall logistic growth rate k using predicted NDVI. Overall, our k-value predictions demonstrate impressive performance, with 50% of predictions exhibiting an absolute error of 0.12 or less, and 75% having an error of 0.24 or less. Finally, we employ Uniform Manifold Approximation and Projection (UMAP) and KNN clustering to identify recovery trends, offering insights into regions with varying rates of recovery. This study pioneers the combined use of tensor regression and ConvLSTM, and introduces the application of UMAP for clustering similar wildfires. This advances predictive ecological modeling and could inform future post-fire vegetation management strategies.