Search papers, labs, and topics across Lattice.
The DCASE 2026 Task 5 introduces Audio-Dependent Question Answering (ADQA), which evaluates whether large audio-language models can derive answers from audio inputs rather than relying on textual information. An innovative Audio-Dependency Filtering (ADF) pipeline was employed to curate a challenging evaluation set of 3000 items across various audio domains, ensuring that only those solvable through audio were included. The top-performing models achieved an accuracy of 58.33%, revealing significant drops in performance on a hidden evaluation split, underscoring the difficulty of the task and the need for robust audio understanding in AI systems.
Audio-language models struggle significantly with hidden evaluation tasks, with accuracy plummeting by nearly 12 percentage points on average, highlighting the challenge of true audio comprehension.
DCASE~2026 Task~5 introduces Audio-Dependent Question Answering (ADQA), which tests whether large audio-language models answer from the audio rather than from textual priors. An Audio-Dependency Filtering (ADF) pipeline combines silent-audio probing, per-option perplexity, a large language model (LLM) commonsense check, and human review to remove items solvable from text alone. The 3000 items that pass form the ADQA-Bench evaluation set, spanning music, speech, and environmental audio. The inaugural edition draws 14 teams and 36 submissions across two tracks defined by total parameter count (up to 100B and under 10B). A Chung-Ang University ensemble of MOSS-Audio-8B-Thinking and Qwen3-Omni-30B reaches the top overall accuracy at \pct{58.33}, and a MOSS-only configuration from the same team leads the sub-10B track at \pct{57.30}. Across the 30 submissions with a comparable development score, evaluation accuracy falls by 11.91 percentage points (pp) on average (median 10.91\,pp) on the hidden evaluation split, which is designed to be harder than the development split. The most common building blocks are: the MOSS-Audio-8B-Thinking backbone (13 of 36 submissions), Low-Rank Adaptation (LoRA) fine-tuning on AudioMCQ-StrongAC, and preference or reinforcement-learning objectives -- Group Relative Policy Optimization (GRPO) in five teams, Group reward-Decoupled Normalization Policy Optimization (GDPO) in two. At test time, prompt engineering is near-universal, and majority or choice-permutation voting is common. Every system misses the same set of 233 evaluation items.
