Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This paper introduces a lightweight episodic memory module for the humanoid robot head Kim, which integrates vector-based semantic retrieval with a language model-controlled dialog system to enhance conversational continuity. The implementation was evaluated through a video-based online study, revealing that the episodic memory significantly improved perceived sociability, trustworthiness, and warmth without increasing perceived disturbance. These results highlight the importance of memory in fostering relational quality in human-robot interactions, suggesting that effective memory systems can enhance social engagement without negative repercussions.
Episodic memory in robots can boost perceived trust and warmth, transforming how we interact with social machines.
Social robots that rely on large language models for conversation are unable to retain information across sessions. This absence of memory violates social expectations, potentially preventing the formation of persistent relationships. This paper presents a lightweight episodic memory module that integrates vector-based semantic retrieval with an LLM-controlled dialog system, deployed on the humanoid robot head Kim. The module employs a hybrid scoring function combining cosine similarity with a memory strength metric to retrieve contextually relevant past interactions and inject them into the generation prompt. The system was evaluated in a within-subjects video-based online study (N = 43) using the Human-Robot Interaction Evaluation Scale (HRIES). Results show that episodic memory significantly increased perceived sociability (d = 0.60, p<.001), with the strongest effects on perceived trustworthiness (d = 0.62) and warmth (d = 0.56). Perceived disturbance remained unchanged (d = 0.00), indicating that the implemented approach to personalized recall did not trigger privacy-related discomfort or uncanny valley effects. These findings suggest that episodic memory serves as a social lubricant in embodied Human-Robot Interaction, enhancing relational quality without eliciting negative affective responses.