# MMA: Multimodal Memory Agent Source: [MMA: Multimodal Memory Agent]() Collection index: [llms.txt]() Data updated: 2026-02-22T06:00:01.441Z Authors: Yihao Lu, Wanru Cheng, Zeyu Zhang, Hao Tang Published: 2026-02-18T14:30:35.000Z ## Abstract Long-horizon multimodal agents depend on external memory; however, similarity-based retrieval often surfaces stale, low-credibility, or conflicting items, which can trigger overconfident errors. We propose Multimodal Memory Agent (MMA), which assigns each retrieved memory item a dynamic reliability score by combining source credibility, temporal decay, and conflict-aware network consensus, and uses this signal to reweight evidence and abstain when support is insufficient. We also introduce MMA-Bench, a programmatically generated benchmark for belief dynamics with controlled speaker reliability and structured text-vision contradictions. Using this framework, we uncover the "Visual Placebo Effect", revealing how RAG-based agents inherit latent visual biases from foundation models. On FEVER, MMA matches baseline accuracy while reducing variance by 35.2% and improving selective utility; on LoCoMo, a safety-oriented configuration improves actionable accuracy and reduces wrong answers; on MMA-Bench, MMA reaches 41.18% Type-B accuracy in Vision mode, while the baseline collapses to 0.0% under the same protocol. Code: https://github.com/AIGeeksGroup/MMA. ## Recorded benchmark results A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. - [Fact Verification on FEVER (Accuracy, Selectivity, and Abstention)](): Fact Verification; FEVER; Accuracy: 59.93. - [Multimodal Fact-checking on MMA-Bench](): Multimodal Fact-checking; MMA-Bench; Core Accuracy: 13.55. - [Task success rate on DMV-Bench J=5 1.0](): Task success rate; DMV-Bench J=5 1.0; Task Success Rate: 47.7. - [Task success rate on DMV-Bench J=10 1.0](): Task success rate; DMV-Bench J=10 1.0; Task Success Rate: 41.7. - [Task success rate on DMV-Bench J=15 1.0](): Task success rate; DMV-Bench J=15 1.0; Task Success Rate: 39.4. - [Task success rate on DMV-Bench MC J=50 1.0](): Task success rate; DMV-Bench MC J=50 1.0; Task Success Rate: 36.9. - [Long-term dialogue memory on LoCoMo (test)](): Long-term dialogue memory; LoCoMo (test); Accuracy: 75.94. ## Original sources and implementation [Original paper]() [PDF]() [Implementation]()