The ICASSP 2026 HumDial Challenge: Benchmarking Human-like Spoken Dialogue Systems in the LLM Era

TL;DR

The paper introduces the HumDial Challenge at ICASSP 2026, a benchmark for evaluating human-like spoken dialogue systems focusing on emotional intelligence and real-time interaction, driven by advancements in LLMs.

Contribution

It establishes the first benchmark for human-like spoken dialogue systems, with a large dataset and evaluation platform for emotional understanding and real-time interaction capabilities.

Findings

Benchmark results demonstrate progress in emotional intelligence.

Evaluation of real-time turn-taking shows system improvements.

Dataset and platform enable standardized assessment of human-like dialogue.

Abstract

Driven by the rapid advancement of Large Language Models (LLMs), particularly Audio-LLMs and Omni-models, spoken dialogue systems have evolved significantly, progressively narrowing the gap between human-machine and human-human interactions. Achieving truly ``human-like'' communication necessitates a dual capability: emotional intelligence to perceive and resonate with users' emotional states, and robust interaction mechanisms to navigate the dynamic, natural flow of conversation, such as real-time turn-taking. Therefore, we launched the first Human-like Spoken Dialogue Systems Challenge (HumDial) at ICASSP 2026 to benchmark these dual capabilities. Anchored by a sizable dataset derived from authentic human conversations, this initiative establishes a fair evaluation platform across two tracks: (1) Emotional Intelligence, targeting long-term emotion understanding and empathetic generation; and (2) Full-Duplex Interaction, systematically evaluating real-time decision-making under `` listening-while-speaking'' conditions. This paper summarizes the dataset, track configurations, and the final results.