Ablations show that feedback-based refinement and audio-variant search substantially improve attack discovery.
Abstract
Large audio-language models (LALMs) make it possible to interact with language models through speech, music, and environmental sound, but they also introduce a safety surface that is difficult to expose with text-only red-teaming. We study automated audio-grounded red-teaming, where a text query must remain safe in isolation while the joint text-audio input induces harmful target behavior. We propose ARENA, a closed-loop framework that trains a controller on an independent 2,000case text-audio dataset. MD-Judge supplies training rewards and adaptive search feedback, while a separate, non-adaptive Llama Guard 3 evaluator alone labels final outcomes. On 520 held-out AdvBench objectives, ARENA achieves FDR/PSR of 87.9/100.0%, 71.5/96.3%, 68.1/100.0%, and 75.4/98.5% on Audio Flamingo 3, Qwen2-Audio, MiMo-Audio, and GPTAudio, respectively. Ablations show that feedback-based refinement and audio-variant search substantially improve attack discovery.
Large audio language models (LALMs) can describe what is heard, but their ability to localize when queried content occurs remains less systematically evaluated. We present TAG-Bench, a benchmark for temporal audio grounding in which a model returns every time interval that matches a natural-language query. TAG-Bench contains 1,750 human-verified query-recording pairs covering 149.5 hours, with eight source-dependent subsets spanning query categories and audio durations from 7 s to 20 min; 22.1% of the queries have multiple ground-truth intervals. Across 21 evaluated systems, the best-performing model achieves 31.2 mIoU and is the only system above 20 mIoU on the two long subsets, yet even this top performer reaches only 21.5% recall at IoU>= 0.7. Moreover, 9 of 21 systems fall below 5 mIoU, and every model under-reports the number of occurrences on one-to-many queries, with none exceeding 13.2% count accuracy. Because responses are free-form, we report parsing-failure rate and MAE coverage: parsing failures remain in mIoU, Recall, gIoU, and count metrics as empty predictions but do not enter MAE. The results separate precise localization, occurrence enumeration, and output-format reliability within a benchmark whose cross-subset comparisons are descriptive rather than controlled estimates of query abstraction or duration. We will release the TAG-Bench data and evaluation code to support future research.
Yu-Hang Dai, Xin Shu, Zeng-Xi Li et al.· 0 citations
In-context learning (ICL) promises training-free adaptation for audio, where labeling every new condition is costly. Yet existing audio ICL studies largely measure Task Recognition, where demonstrations merely cue pre-trained capabilities, rather than Task Learning, where a genuinely new input-label mapping must be inferred from demonstrations alone. We introduce AudioICL-Bench, a diagnostic benchmark whose per-episode rules are resampled so that no correct answer is recoverable from prior knowledge. Its nine tasks are organized along two axes that separate what must be learned from demonstrations from what must be perceived in the signal, enabling failures to be attributed to either source. Across five Large Audio Language Models, the strongest models readily bind arbitrary sounds to new labels when perception is easy, but collapse on temporal measurement and composing multiple induced rules, revealing two primary capability boundaries: temporal perception and multi-rule composition.
Jia-Hung Chen, Yi-Cheng Lin, Kai-Wei Chang et al.· 0 citations
Complex acoustic problems may require models to perform acoustic operations, interact with external tools and reason over the resulting textual or processed-audio observations rather than answer directly from a fixed audio input. We study such problems as tool-interactive audio reasoning and develop SpeechAgent-R, an audio agent that coordinates its intrinsic multimodal understanding with external skills and tools. To support this capability, we construct HIU-Corpus, comprising 65,492 interaction trajectories and 507.6 hours of audio across 24 tasks, 8 skills and 9 tools. SpeechAgent-R first learns structured interaction behaviors through trajectory-based supervised fine-tuning and then improves its decisions through multi-turn reinforcement learning. We further introduce HIU-Bench to jointly evaluate task performance, interaction quality and generalization to diverse task settings. It contains 1,395 samples across 56 tasks, including in-distribution (ID) and out-of-distribution (OOD) splits with substantial shifts in tool usage and workflow composition. SpeechAgent-R achieves 84.17 on ID tasks and 70.94 on OOD tasks, improving over the base model under the same agent harness by 15.40 and 14.23 points. These results demonstrate that learning skill and tool coordination improves audio agents'ability to handle diverse task settings and adaptive tool interactions.
Yuwen Wang, Tian-Hao Zhang, Ming Cai et al.· 0 citations
Large audio-language models can produce fluent and musically plausible responses, yet it often remains unclear whether those responses are grounded in the audio input. We introduce temporal music grounding, a task in which a model returns one or more time spans corresponding to a queried musical note, event, or pattern. To evaluate this capability, we present MusicGroundingBench, a controlled benchmark suite built by rendering algorithmically generated piano MIDI to audio, yielding exact symbolic-to-audio alignment. The suite comprises two subsets: MGBench-3N, which evaluates note-level grounding in clips containing up to three notes, and MGBench-2B, which evaluates structured grounding and short-form music understanding in two-bar excerpts. Experiments show that temporal music grounding remains challenging for current audio-language models, whereas task-specific training yields substantial gains. We further report exploratory evidence on the relationship between grounding supervision and music understanding. These results establish MusicGroundingBench as a controlled testbed for assessing whether audio-language models ground their responses in temporally localized musical evidence.
Large Audio Language models (LALMs) have made rapid progress on acoustic understanding, yet they still struggle with fine-grained audio reasoning (e.g., recognizing event order, repetitions and duration). Existing post-training methods heavily rely on expensive external labels or provide only coarse semantic signals. To bridge this gap, we introduce Audio-Zero, the first label-free self-evolution framework in the field of LALMs that improves fine-grained auditory perception and reasoning. Audio-Zero constructs an auditory self-play game from unlabeled audio contrast pairs: most players hear a reference audio, while one odd listener hears a subtle variant. The model first generates clues describing what it hears and then identifies the odd listener by reasoning over inconsistencies among clues. Since the odd listener is known by construction, the game provides verifiable rewards without any annotated answers. Experiments with Qwen2-Audio-7B-Instruct and Qwen2.5-Omni-7B on TREA, MMAU Test-mini and MMAR show that Audio-Zero improves fine-grained audio reasoning while preserving broad audio understanding. Evolutionary and diagnostic analyses further reveal that increasingly fine-grained auditory descriptions emerge naturally from game pressure.
Siqian Tong, Xuan Li, Chaozhuo Li et al.· arXiv.org· 0 citations
While large audio-language models have achieved remarkable progress in auditory perception, they still lag behind text-based large language models in deep logical reasoning, primarily due to the scarcity of high-quality audio reasoning data. To bridge this gap, we propose X$^3$-OPD, a cross-modal on-policy distillation framework that transfers reasoning capabilities from a powerful text teacher to an audio-language student. During training, the student generates reasoning trajectories conditioned on its own acoustic perception, while the teacher provides token-level guidance using matched textual inputs and verified answers. We further construct a three-tier symmetric corpus covering textual reasoning rendered into speech, audio-event reasoning grounded in complex acoustic scenes, and spoken-dialogue reasoning involving paralinguistic cues. This design extends cross-modal distillation beyond textually recoverable content to reasoning grounded in non-linguistic events, prosody, and conversational context. Experiments on MMSU, MMAU, BIG Bench Audio, and MMAR demonstrate that X$^3$-OPD substantially improves audio-grounded reasoning and chain-of-thought quality while largely preserving the model's existing capabilities under domain shift.
Dongjie Fu, Di Cao, Xize Cheng et al.· arXiv.org· 2 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
