Multi-Source Evidence Fusion for Audio Question Answering
This addresses the need for verifiable reasoning in audio language models for researchers and practitioners, though it appears incremental as it builds on existing models and tools.
The paper tackles the problem of opaque reasoning in audio question answering systems by developing a multi-source ensemble pipeline that grounds inferences in explicit, reliability-tagged evidence, resulting in a system that ranked first in the Interspeech 2026 Audio Reasoning Challenge with a wide margin in reasoning quality.
Large audio language models (LALMs) can answer questions about speech, music, and environmental sounds, yet their internal reasoning is largely opaque and difficult to validate. We describe TalTech's solution to the Agent Track of the Interspeech 2026 Audio Reasoning Challenge, in which systems are evaluated on reasoning process quality, specifically the factual accuracy, logical soundness, and completeness of their reasoning chains. Our multi-source ensemble pipeline uses two LALMs that generate independent observations, while a separate text-only reasoning model cross-checks these against outputs from 25 acoustic tools organized into reliability tiers. By grounding every inference step in explicit, reliability-tagged evidence, the system produces dense, verifiable reasoning chains. Our system ranked first in the challenge, outperforming all competing systems by a wide margin in challenge's reasoning quality metric.