The proposed symbiotic architecture employs an injector module that writes audio-conditioned vectors directly into the target LLM's short-term memory, enabling the LLM to behave as an audio language model (ALM) and improves the scalability of ALMs.
Abstract
This paper proposes an architecture for equipping large language models (LLMs) with audio-understanding capabilities without fine-tuning their weights. The proposed symbiotic architecture employs an injector module that writes audio-conditioned vectors directly into the target LLM's short-term memory, i.e., the key-value (KV) cache, enabling the LLM to behave as an audio language model (ALM). The architectural advantages are twofold. First, it improves the scalability of ALMs: because the proposed method bypasses the LLM during audio injection, the injection cost is governed by the injector width rather than the backbone width, and can therefore scale more slowly than the cost of full-backbone prefilling. Second, since the training scheme does not update the LLM weights, the original capabilities of the LLM are preserved without the risk of degradation from fine-tuning. The effectiveness of the proposed method is evaluated on both audio-understanding tasks (automatic speech recognition, audio question answering, and acoustic scene classification) and text-only tasks. We confirm that, while activating fewer parameters during audio prefilling, our architecture outperforms the conventional method with a frozen LLM and approaches the performance of a fine-tuned ALM, all while preserving the backbone LLM's original text-only task performance by construction.
Mizar, a 159.3M-parameter ALM, is introduced, a recipe that brings together architecture, data, and three-stage training to build Mizar, a 159.3M-parameter ALM that surpasses the previous best-performing ALM below 200M parameters on all three benchmarks.
This paper describes the system for Task~2 of the second Multilingual Conversational Speech Language Model (MLC-SLM) Challenge, which adapt Qwen3-Omni-30B-A3B-Instruct with a segment-evidence-aware data and post-training pipeline and obtains 90.92% accuracy on the final official evaluation set.
H. Le, L. Nguyen, Minh Tri Dao· 2nd Workshop on Multilingual...· 1 citation
X-AuT, a progressive framework that selects layer combinations through short behavioral probes and restores the pruned model through representation alignment, cross-scale distillation, scheduled student-policy supervision, and LoRA finetuning, is introduced.
Hao-Jun Zhang, Yi Zou, Min Chen et al.· 0 citations
This work presents TEMPO (Temporally-grounded Multi-task Post-training), the first unified model to handle audio, speech, and music timestamping tasks and introduces the first application of reinforcement learning to unified audio timestamping, using GRPO with verifiable temporal rewards that directly optimize the eval...
Apoorva Kulkarni, Kaousheik Jayakumar, Sreyan Ghosh et al.· 2 citations
The design of transformer-based Large Language Models (LLMs) is being radically changed through new architectures that are able to overcome scalability limitations of previous designs, including Mixture-of-Experts (MoE), Multi-Head Latent Attention (MLA), and Multi-Token Prediction (MTP). As an open-weighted model rele...
Yassine Zouhdi, B. Hdioud· EPJ Web of Conferences· 0 citations
This paper investigates how to reduce the amount of transmitted data during interactions with MLLMs while preserving their multimodal understanding performance, and proposes a token communication framework tailored to MLLMs.
Jing-Kai Ying, Zhi-Jin Qin, Yuan Shen et al.· 0 citations
Related blog posts
MIT News · Artificial Intelligence· news.mit.eduNov 19, 2026
An MIT and Dana-Farber Institute-led team used large language models to examine nearly 4,000 studies and hundreds of code repositories, revealing gaps in the information needed to scrutinize and reuse published research.
By studying the computational foundations of language learning, Associate Professor Jacob Andreas builds intelligent systems that can communicate effectively with humans.