Real-Time Conversational Agents (RTCA) Workshop @ NeurIPS 2026
Sydney, Australia — 11 or 12 December 2026
Website: https://rtcaneurips26.github.io/
Submission portal (OpenReview): https://openreview.net/group?id=NeurIPS.cc/2026/Workshop/RTCA

*** Submission deadline: 29 August 2026 (Anywhere on Earth) ***

We are pleased to share the Call for Papers and Demos for the inaugural Real-Time Conversational Agents (RTCA) Workshop, held in conjunction with NeurIPS 2026 in Sydney, Australia.

OVERVIEW

Conversational AI has moved from text chat into the real world — voice modes, embodied avatars, and agents that share our screens and tools. To feel natural, these systems must operate in real time: streaming speech, video, and language while continuously listening, watching, and re-planning. This is fundamentally harder than offline generation — latency, turn-taking, backchannels, interruptions, and cross-modal alignment all become first-class problems that the offline-generation paradigm largely sidesteps. Recent progress on full-duplex speech-language models, real-time talking-head and avatar generation, low-latency speech synthesis, and streaming ASR shows that interactive multimodal agents are now technically feasible, but the field still lacks shared benchmarks, vocabulary, and methodology for interactional naturalness as distinct from per-utterance quality.

RTCA brings together researchers across speech, vision, language, HCI, social-signal processing, and ML systems around three intertwined questions:

Real-time generation under hard latency budgets — producing high-quality speech, video, and language in a streaming or full-duplex fashion.
Naturalness in interaction — what makes an agent feel like a conversational partner (prosody, gaze, timing, grounding, expressivity, turn-taking).
Evaluation of live systems — measuring naturalness, responsiveness, and conversational quality where standard offline metrics fall short.

TOPICS OF INTEREST (non-exhaustive)

Streaming/low-latency speech synthesis, ASR, and full-duplex audio-language models
Real-time talking-head, avatar, and embodied video generation; lip-sync, gaze, expressivity under streaming
Streaming language models; incremental and speculative decoding for dialogue
Turn-taking, backchanneling, interruption handling, and floor management
Multimodal alignment under latency and partial-observation constraints
Prosody, emotion, and paralinguistic generation in interactive settings
Memory, grounding, and tool use during live conversation
Evaluation of naturalness: perceptual studies, turn-taking metrics, perceived latency, interactive Turing-style tests
Datasets and benchmarks for interactive (not offline) evaluation
Efficient inference, on-device deployment, and the systems-quality trade-off
Safety, identity, and trust in real-time agents (deepfakes, persuasion, consent)

Position papers, critiques of current evaluation practice, and reproducibility studies are also welcome.

SUBMISSION TYPES

Full papers (up to 8 pages) — original contributions; may be presented as posters and/or contributed talks.
Short papers (up to 4 pages) — work in progress, focused contributions, or position papers.
Demo papers (extended abstract or up to 2 pages) — required for the on-stage Conversational Agents Showcase.

All submissions must use the NeurIPS 2026 style file and be formatted for double-blind review. Page limits exclude references and appendices. Papers must be submitted in PDF format via OpenReview. The workshop is non-archival; authors retain the right to publish elsewhere.

IMPORTANT DATES (End of day, Anywhere on Earth)

Call for papers opens: 18 July 2026
Submission deadline (papers and demos): 29 August 2026
Author notification: 29 September 2026
Workshop date: 11 or 12 December 2026

CONFIRMED INVITED SPEAKERS

Dimitris Samaras — Stony Brook University, USA
Evonne Ng — Meta Reality Labs / UC Berkeley, USA

ORGANISERS

Niki Foteinopoulou — Tavus, United Kingdom
Alessandro Conti — Tavus, Italy
Jack Saunders — Tavus, United Kingdom
Oya Celiktutan — King's College London, United Kingdom
Cigdem Beyan — University of Verona, Italy
Ioannis Patras — Queen Mary University of London, United Kingdom

CONTACT

For more information, visit https://rtcaneurips26.github.io/ or contact us at rtca-workshop@googlegroups.com.

We look forward to your contributions.