An Evaluation of Interleaved Instruction Tuning on Semantic Reasoning Performance in an Audio MLLM
This work addresses the problem of limited modality integration in MLLMs for audio-based semantic reasoning, though it is incremental with trade-offs.
The study investigated interleaved instruction tuning in an audio MLLM to enhance semantic reasoning, finding that fine-tuning improved performance on synonym and hypernym tasks but reduced audio labeling ability.
Standard training for Multi-modal Large Language Models (MLLMs) involves concatenating non-textual information, like vision or audio, with a text prompt. This approach may not encourage deep integration of modalities, limiting the model's ability to leverage the core language model's reasoning capabilities. This work examined the impact of interleaved instruction tuning in an audio MLLM, where audio tokens are interleaved within the prompt. Using the Listen, Think, and Understand (LTU) model as a testbed, we conduct an experiment using the Synonym and Hypernym Audio Reasoning Dataset (SHARD), our newly created reasoning benchmark for audio-based semantic reasoning focusing on synonym and hypernym recognition. Our findings show that while even zero-shot interleaved prompting improves performance on our reasoning tasks, a small amount of fine-tuning using interleaved training prompts improves the results further, however, at the expense of the MLLM's audio labeling ability.