Enabling Conversational Behavior Reasoning Capabilities in Full-Duplex Speech

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Pan, Shuchang, Banerjee, Siddharth, Hebbar, Dhruv, Patel, Siddhant, Gupta, Akshaj, Cheng, Kan Jen, Kim, Hanjo, Li, Zeyi Austin, Ma, Martin Q., Li, Tingle, Anumanchipalli, Gopala, Lian, Jiachen
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866917169698177024
author Pan, Shuchang
Banerjee, Siddharth
Hebbar, Dhruv
Patel, Siddhant
Gupta, Akshaj
Cheng, Kan Jen
Kim, Hanjo
Li, Zeyi Austin
Ma, Martin Q.
Li, Tingle
Anumanchipalli, Gopala
Lian, Jiachen
author_facet Pan, Shuchang
Banerjee, Siddharth
Hebbar, Dhruv
Patel, Siddhant
Gupta, Akshaj
Cheng, Kan Jen
Kim, Hanjo
Li, Zeyi Austin
Ma, Martin Q.
Li, Tingle
Anumanchipalli, Gopala
Lian, Jiachen
contents Human conversation is organized by an implicit chain of thoughts that manifests as timed speech acts. Capturing this causal pathway is key to building natural full-duplex interactive systems. We introduce a framework that enables reasoning over conversational behaviors by modeling this process as causal inference within a Graph-of-Thoughts (GoT). Our approach formalizes the intent-to-action pathway with a hierarchical labeling scheme, predicting high-level communicative intents and low-level speech acts to learn their causal and temporal dependencies. To train this system, we develop a hybrid corpus that pairs controllable, event-rich simulations with human-annotated rationales and real conversational speech. The GoT framework structures streaming predictions as an evolving graph, enabling a multimodal transformer to forecast the next speech act, generate concise justifications for its decisions, and dynamically refine its reasoning. Experiments on both synthetic and real duplex dialogues show that the framework delivers robust behavior detection, produces interpretable reasoning chains, and establishes a foundation for benchmarking conversational reasoning in full duplex spoken dialogue systems.
format Preprint
id arxiv_https___arxiv_org_abs_2512_21706
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Enabling Conversational Behavior Reasoning Capabilities in Full-Duplex Speech
Pan, Shuchang
Banerjee, Siddharth
Hebbar, Dhruv
Patel, Siddhant
Gupta, Akshaj
Cheng, Kan Jen
Kim, Hanjo
Li, Zeyi Austin
Ma, Martin Q.
Li, Tingle
Anumanchipalli, Gopala
Lian, Jiachen
Computation and Language
Artificial Intelligence
Human conversation is organized by an implicit chain of thoughts that manifests as timed speech acts. Capturing this causal pathway is key to building natural full-duplex interactive systems. We introduce a framework that enables reasoning over conversational behaviors by modeling this process as causal inference within a Graph-of-Thoughts (GoT). Our approach formalizes the intent-to-action pathway with a hierarchical labeling scheme, predicting high-level communicative intents and low-level speech acts to learn their causal and temporal dependencies. To train this system, we develop a hybrid corpus that pairs controllable, event-rich simulations with human-annotated rationales and real conversational speech. The GoT framework structures streaming predictions as an evolving graph, enabling a multimodal transformer to forecast the next speech act, generate concise justifications for its decisions, and dynamically refine its reasoning. Experiments on both synthetic and real duplex dialogues show that the framework delivers robust behavior detection, produces interpretable reasoning chains, and establishes a foundation for benchmarking conversational reasoning in full duplex spoken dialogue systems.
title Enabling Conversational Behavior Reasoning Capabilities in Full-Duplex Speech
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2512.21706