Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
October 20, 2025Open Access

FindingDory: A Benchmark to Evaluate Memory in Embodied Agents

View Full Paper
Ask AI
Bookmark
Share

Authors

KYKarmesh YadavYAYusuf AliGGGunshi Gupta

Discussion

Loading...

Member takes

Overview

This benchmark assesses memory integration in embodied agents across 60 tasks, suggesting improvements in reasoning and navigation performance.

Key Points

  • The benchmark evaluates memory capabilities in embodied tasks requiring sustained engagement and awareness.
  • Sixty tasks designed to put memory integration and long-term experience to the test are included.
  • Integrating state-of-the-art VLMs with navigation policies shows varying success in memory-intensive tasks.
  • Enhancing memory mechanisms in embodied agents can significantly improve performance in long-range control.

Cite This Study

Yadav et al. (2025) studied this question.

synapsesocial.com/papers/68f6196ee0bbbc94fac36238https://doi.org/10.48550/arxiv.2506.15635
View Full Paper
Ask AI
Bookmark
Share

Also Consider

Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context:

  1. 1VideoAgent: A Memory-augmented Multimodal Agent for Video Understanding2024 · 2 citations
  2. 2Human-like memory empowers embodied robots for long-term object navigation2026
  3. 3Memory-Maze: Scenario Driven Benchmark and Visual Language Navigation Model for Guiding Blind People2024
  4. 4EmbRACE-3K: Embodied Reasoning and Action in Complex Environments2025
  5. 5Dream to Recall: Imagination-Guided Experience Retrieval for Memory-Persistent Vision-and-Language Navigation2026