CareAgent: In Our Next Session, Let Me Actively Care for YouDownload PDF

Anonymous

16 Feb 2024ACL ARR 2024 February Blind SubmissionReaders: Everyone
Abstract: Despite the fact that the large language models (LLMs) facilitate conversations with rationalization and knowledge, they are still restricted to the passive response mechanism that relies on user instructions, which are hard to actively care for the user. To realize the actively caring ability, we propose an active conversational agent (ACA) named CareAgent which creates new session to approach user potential interest. Specifically, inspired by the Jung’s theory of psychological types and the active exploration mechanism of agent in environment, we designed three components to support the goal of caring for user. The Character Extractor (CE) obtains the personality through Myers-Briggs Type Indicator (MBTI) and attributes for character descriptions; the Memory Reconstructor (MR) achieves multi-topic summaries based on multi-clue branching for complete memories; the Decision Adapter (DA) selects the best topic summary as the background memory and adapts the agent intention to control the scenario of new session. The results of experiments demonstrated that CareAgent was able to maintain reliability in character understanding and extract complete multi-topic summaries from conversational history. Evaluators also believed that this agent enhanced the actively caring and personification level in new session.
Paper Type: long
Research Area: Dialogue and Interactive Systems
Contribution Types: Model analysis & interpretability, NLP engineering experiment
Languages Studied: English
Preprint Status: There is no non-anonymous preprint and we do not intend to release one.
A1: yes
A1 Elaboration For Yes Or No: Section 6
A2: yes
A2 Elaboration For Yes Or No: Section 7
A3: yes
A3 Elaboration For Yes Or No: Abstract and Section 1
B: yes
B1: yes
B1 Elaboration For Yes Or No: Section 3; Sectioin 4.1; Abstract Description of 16 Personalities in Appendix C; Appendix G
B2: yes
B2 Elaboration For Yes Or No: Sectioin 4.1; Appendix G
B3: yes
B3 Elaboration For Yes Or No: Sectioin 4.1; Section 7; Appendix G
B4: yes
B4 Elaboration For Yes Or No: Section 7
B5: yes
B5 Elaboration For Yes Or No: Abstract Description of 16 Personalities in Appendix C; Appendix G
B6: yes
B6 Elaboration For Yes Or No: Sectioin 4.1; Appendix G
C: yes
C1: yes
C1 Elaboration For Yes Or No: Section 4.2
C2: yes
C2 Elaboration For Yes Or No: Section 4.2
C3: yes
C3 Elaboration For Yes Or No: Section 4.3; Section 4.4
C4: yes
C4 Elaboration For Yes Or No: Appendix G
D: yes
D1: yes
D1 Elaboration For Yes Or No: Appendix G
D2: yes
D2 Elaboration For Yes Or No: Appendix G
D3: yes
D3 Elaboration For Yes Or No: Appendix G
D4: yes
D4 Elaboration For Yes Or No: Appendix G; Section 7
D5: yes
D5 Elaboration For Yes Or No: Appendix G
E: yes
E1: yes
E1 Elaboration For Yes Or No: The ChatGPT API was used for the experiment in Section 4
0 Replies

Loading

OpenReview is a long-term project to advance science through improved peer review with legal nonprofit status. We gratefully acknowledge the support of the OpenReview Sponsors. © 2025 OpenReview