Robust ASR

We describe the Itgan systems for the three ASR subtasks of NADI 2026, namely robust country-level ASR (1.1), mixed-dialect ASR (1.2), and Tunisian code-switched ASR (1.3). All three share one recipe, Whisper adapted with LoRA on consumer GPUs, and each was carried by a different addition to it. On 1.1, where the dialect label is given at test time, per-dialect specialists continued from a pooled adapter gave the largest gain, and the submitted system reached 57.1% country-average WER. A post-evaluation linear probe on frozen encoder features routes utterances without the label and recovers 44% of what oracle routing gives. On 1.2 the choice of base model mattered more than adapter capacity, and system combination helped only once we added a decorrelated member, reaching 46.7% WER. On 1.3 our system placed second at 14.49% WER with the lowest CER among the leading submissions, 5.38%. Its last 0.60 WER points came without further training, mostly from an exact weight-space average of independently trained runs, with ROVER voting adding the remainder. Every comparison carries a paired-bootstrap test, and we report eight directions that did not work.

Current audio-visual speech recognition (AVSR) benchmarks, like LRS3, rely heavily on clean, scripted and rehearsed speech. They fail to reflect the complexity of natural conversation, which involves overlapping speech, spontaneous turn-taking, unscripted vocabulary and variable acoustic conditions. To shift the field toward realistic dialogue, we introduce Candor-LR, a conversational benchmark derived from the CANDOR corpus of 1,656 natural dyadic videoconferences. Our custom data preparation pipeline yields 713.5, 10.1, and 60.1 hours of training, validation, and test data, respectively. Evaluating pretrained AVSR models on Candor-LR reveals that audio-only accuracy drops sharply compared to LRS3, but visual cues compensate effectively, driving much larger performance gains on Candor-LR than on LRS3. Furthermore, training on this corpus significantly improves cross-domain robustness under both clean and noisy conditions, as its realistic conversational data captures broader audio-video features. We open-source our pipeline to ensure reproducibility, establishing Candor-LR as a challenging benchmark for conversational AVSR.

Robust speech-to-text translation systems should perform reliably across diverse acoustic conditions, yet practical pipelines lack controllable tools for systematic environment exploration. Large speech models remain sensitive to unseen acoustic conditions, as training data rarely cover the full range of real environments.We present NOPEHYPE, a structured training workflow that combines a controllable environment simulator, coverage-optimal environment reduction on Power Spectral Density (PSD) templates, and a small, interpretable hyperparameter search over simulator knobs. We show that simulator-generated noise achieves performance comparable to balanced realnoise training across Whisper and SeamlessM4T models, provide principled environment prototype sets, and identify practical default simulator configurations from a structured 27-run hyperparameter sweep.

ASR systems sometimes produce fluent text that is unrelated to the speech they receive. We view these hallucinations as one possible consequence of a broader grounding failure, in which the transcript is no longer adequately guided by the audio. To understand where this failure becomes possible, we study two independently trained Conformer-Large recognizers - one CTC and one RNN-T - under environmental degradation and speaker-background shift. In both models, the final encoder stage emerges as a critical boundary: bypassing the final block causes divergence on nearly every utterance, whereas bypassing middle blocks has little effect. At this same stage, the representations become more compact, text becomes readable by the trained decoder, and grapheme information becomes explicit. Importantly, the intervention produces garbled or repetitive output rather than fluent fabrication. Our result therefore identifies a mechanistic precondition for hallucination - the failure to produce adequately grounded output - not the complete origin of naturally occurring hallucinations. Together, the results reveal a consistent terminal-stage dependency for grounded recognition across two decoder families and multiple distribution shifts.

Humans are often considered to be the best listeners and seen as the upper-bound performance of automatic speech recognition (ASR) systems. We present a preliminary comparison of the performances of state-of-the-art ASR systems and Dutch native listeners on the recognition of "diverse" speech, specifically Dutch child and older adults' speech and Flemish. Google Telephony outperformed the other ASR systems. Importantly, the ASR systems showed similar performance to the listeners, and in specific cases even outperformed them. Slight performance differences between the listeners and ASR systems were found related to speaker's age and regional accents and utterance length. Future research should focus on making ASR systems more robust to acoustic variability related to aging and regional accents. A comparison of ASR recognition performances on the test stimuli and the full Jasmin-CGN test sets showed the influence of the specific test sets on the conclusions regarding benchmarking human and ASR performance.

Audio-Visual Speech Recognition takes two input modalities, acoustic and visual streams, where visual information from lip movements aids recognition when audio is noisy. Recently, LLM-based AVSR models have emerged as a promising paradigm by connecting pre-trained audio-visual encoders to an LLM, achieving strong results in clean conditions. However, these models are predominantly optimized for clean acoustic conditions, with limited attention to making the LLM backbone robust to noise. No explicit mechanism is employed to produce stable representations under corrupted audio, leading to performance degradation in noisy environments. To address this, we propose VIB-AVSR, which integrates Variational Information Bottleneck layers at targeted positions within the LLM backbone to regularize representations. VIB-AVSR reduces degradation under noisy conditions across multiple SNR levels and noise types, without requiring architectural modifications or additional training data.

Automatic Speech Recognition systems are notoriously both sensitive to adversarial and benign perturbations. While this has been repeatedly demonstrated using reference datasets, detecting such behaviors in deployed systems is incredibly challenging, due to the absence of oracle knowledge of the true transcription. We demonstrate that employing a certification-inspired mechanism can significantly decrease WER, increase recall, and decrease the Spearman correlation between confidence and WER. We achieve this through a dual-gate diagnostic pipeline: a Two-Sided Atomic Audit that accumulates statistical wealth to certify both token existence and adversarial exclusion, and a Rank-Based Tournament that selects the winning sequence. Our evaluations across four diverse architectures demonstrate up to a 55% relative reduction in Word Error Rate, while also providing granular word- and sentence-level certifications to enhance acoustic security.

Modern Automatic Speech Recognition (ASR) systems have made remarkable progress on standard benchmarks, yet performance gaps have emerged under real-world distribution shifts, caused by recording conditions, accents, speech impairments, and noise. Existing datasets and benchmarks typically isolate these factors, which overlooks their co-occurrence in real-world applications. In this paper, we argue that model robustness can be treated as a dynamic capability that continually develops, and we introduce MoDiCoL, a Modular Diagnostic Continual Learning dataset designed for controlled analysis of linguistic content, speaker characteristics, and acoustic environments. Furthermore, we propose a real-world-inspired continual learning curriculum to simulate incremental updates and study how robustness is acquired, transferred, and forgotten. We evaluate three continual learning strategies and provide detailed insights into robustness under evolving conditions.

Despite rapid advances in automatic speech recognition (ASR) and large audio-language models, robust recognition in real-world environments remains limited by an "acoustic robustness bottleneck": models often lose acoustic grounding and produce omissions or hallucinations under severe, compositional distortions. We propose Mega-ASR, a unified ASR-in-the-wild framework that combines scalable compound-data construction with progressive acoustic-to-semantic optimization. We introduce Voices-in-the-Wild-2M, covering 7 classic acoustic phenomena and 54 physically plausible compound scenarios, and train Mega-ASR with Acoustic-to-Semantic Progressive Supervised Fine-Tuning and Dual-Granularity WER-Gated Policy Optimization. Extensive experiments demonstrate that Mega-ASR achieves significant advantages over prior state-of-the-art systems on adverse-condition ASR benchmarks (45.69% vs. 54.01% on VOiCES R4-B-F, and 21.49% vs. 29.34% on NOIZEUS Sta-0). On complex compositional acoustic scenarios, Mega-ASR further delivers over 30% relative WER reduction against strong open- and closed-source baselines, establishing a scalable paradigm for robust ASR in-the-wild.

Fine-tuning multilingual ASR models like Whisper for low-resource languages often improves read speech but degrades spontaneous audio performance. To diagnose this mismatch, we introduce Vividh-ASR, a complexity-stratified benchmark for Hindi and Malayalam across four tiers: studio, broadcast, spontaneous, and synthetic noise. Through a controlled study of learning-rate timing and curriculum ordering, we find that early large parameter updates improve global WER by 12 absolute points, while a hard-to-easy curriculum adds gains for spontaneous speech. These findings motivate reverse multi-stage fine-tuning (R-MFT), a training recipe that enables a parameter-efficient 244M Whisper model to match or exceed conventionally fine-tuned 769M counterparts. Representational analysis via CKA and SVD reveals effective schedules concentrate adaptation in the decoder, preserving the pre-trained encoder's acoustic geometry. We release the benchmark and models.

LLMs can estimate Hospital Anxiety and Depression Scale (HADS) scores from speech in a zero-shot manner, but clinical deployment requires reliability across three dimensions: intra-model consistency, ASR robustness, and evidence faithfulness. We evaluate three LLMs (Phi-4, Gemma-2-9B, and Llama-3.1-8B) on 111 English-speaking participants using ground-truth transcripts and three Whisper ASR variants (Large, Medium, Small), with three independent runs per model-condition pair. We find that (i) Phi-4 and Gemma-2-9B achieve excellent intra-model consistency (ICC > 0.89) with minimal degradation under ASR; (ii) Llama-3.1-8B shows ASR-fragile consistency, with ICC dropping from 0.82 to 0.36 at 10% WER; (iii) predictive validity is largely preserved under ASR for robust models; and (iv) keyword groundedness exceeds 93% for Phi-4 and Gemma-2-9B but falls to 77-81% for Llama-3.1-8B. Inter-model keyword agreement is far lower than score-level agreement, revealing a score-evidence dissociation with implications for clinical interpretability.

Recent large language models (LLMs) show strong speech recognition and translation capabilities for high-resource languages. However, African languages remain dramatically underrepresented in benchmarks, limiting their practical use in low-resource settings. While early benchmarks tested African languages and accents, they lacked exhaustive real-world noise and granular domain evaluations. We present AfriVox-v2, a comprehensive benchmark designed to test speech models under realistic African deployment conditions. AfriVox-v2 introduces "in the wild" unscripted audio for all supported languages. We also introduce strict domain verticalization, evaluating model accuracy across ten sectors including government, finance, health, and agriculture and conducting targeted tests on numbers and named entities. Finally, we benchmark a new generation of speech models, including Sahara-v2, Gemini 3 Flash, and the Omnilingual CTC models. Our results expose the true generalization gap of modern speech models in specialized, noisy African contexts and provide a reliable blueprint for developers building localized voice AI.

ASR systems based on self-supervised acoustic pretraining and CTC fine-tuning achieve strong performance on native speech but remain sensitive to accent variability. We investigate supervised contrastive learning (SupCon) as a lightweight, accent-invariant auxiliary objective for CTC fine-tuning. An utterance-level contrastive loss regularizes encoder representations without architectural modification or explicit accent supervision. Experiments on the L2-ARCTIC benchmark show consistent WER reductions across multiple pretrained encoders, with up to 25 -- 29% relative reduction under unseen-accent evaluation. Analysis using within-transcript cosine dispersion indicates that SupCon promotes more compact and stable representation geometry under accent variability. Overall, SupCon provides an effective and model-agnostic regularization strategy for improving accent robustness.

We introduce LRS-VoxMM, an in-the-wild benchmark for audio-visual speech recognition (AVSR). The benchmark is derived from VoxMM, a dataset of diverse real-world spoken conversations with human-annotated transcriptions. We select AVSR-suitable samples and preprocess them in an LRS-style format for direct use in existing AVSR pipelines. Compared with commonly used benchmarks, LRS-VoxMM covers a more diverse range of scenarios and acoustic conditions. We also release distorted evaluation sets with additive noise, reverberation, and bandwidth limitation to support evaluation under severe acoustic degradation. Experimental results show that LRS-VoxMM is considerably harder than LRS3 and that the contribution of visual information becomes more evident as the audio signal degrades. LRS-VoxMM supports more realistic AVSR benchmarking and encourages further research on the role of visual information in challenging real-world conditions.

As pretrained large language models replace task-specific decoders in speech recognition, a critical question arises: do their text-derived priors make recognition fairer or more biased across demographic groups? We evaluate nine models spanning three architectural generations (CTC with no language model, encoder-decoder with an implicit LM, and LLM-based with an explicit pretrained decoder) on about 43,000 utterances across five demographic axes (ethnicity, accent, gender, age, first language) using Common Voice 24 and Meta's Fair-Speech, a controlled-prompt dataset that eliminates vocabulary confounds. On clean audio, three findings challenge assumptions: LLM decoders do not amplify racial bias (Granite-8B has the best ethnicity fairness, max/min WER = 2.28); Whisper exhibits pathological hallucination on Indian-accented speech with a non-monotonic insertion-rate spike to 9.62% at large-v3; and audio compression predicts accent fairness more than LLM scale. We then stress-test these findings under 12 acoustic degradation conditions (noise, reverberation, silence injection, chunk masking) across both datasets, totaling 216 inference runs. Severe degradation paradoxically compresses fairness gaps as all groups converge to high WER, but silence injection amplifies Whisper's accent bias up to 4.64x by triggering demographic-selective hallucination. Under masking, Whisper enters catastrophic repetition loops (86% of 51,797 insertions) while explicit-LLM decoders produce 38x fewer insertions with near-zero repetition; high-compression audio encoding (Q-former) reintroduces repetition pathology even in LLM decoders. These results suggest that audio encoder design, not LLM scaling, is the primary lever for equitable and robust speech recognition.

Integrating large language models (LLMs) into automatic speech recognition (ASR) has become a mainstream paradigm in recent years. Although existing LLM-based ASR models demonstrate impressive performance on public benchmarks, their training remains predominantly data-driven, leaving key practical challenges insufficiently addressed -- particularly limited downward scalability in resource-constrained deployments and hallucinations under acoustically challenging conditions. To address these issues, we present NIM4-ASR, a production-oriented LLM-based ASR framework optimized for both efficiency and robustness. Grounded in a principled delineation of functional roles between the encoder and the LLM, we redesign the multi-stage training paradigm to align each module with its intended capability boundary. Specifically, we reformulate the pre-training architecture and objective to mitigate the modality gap and improve parameter efficiency; introduce an iterative asynchronous SFT stage to preserve acoustic fidelity and constrain representation drift; and design an ASR-specialized reinforcement learning stage to further enhance recognition quality and robustness. We additionally incorporate a suite of production-oriented optimizations, including robustness under noisy and silent conditions, real-time streaming inference, and hotword customization via retrieval-augmented generation (RAG). Experiments show that NIM4-ASR achieves state-of-the-art performance on multiple public benchmarks with merely 2.3B parameters, while substantially outperforming larger-scale competitors on internal benchmarks -- particularly in entity-intensive real-world scenarios. NIM4-ASR further supports million-scale hotword customization via RAG with sub-millisecond retrieval latency, enabling efficient adaptation to emerging entities and personalized user requirements.

Automatic speech recognition (ASR) has been extensively studied on neutral and stationary speech, yet its robustness under post-exercise physiological shift remains underexplored. Compared with resting speech, post-exercise speech often contains micro-breaths, non-semantic pauses, unstable phonation, and repetitions caused by reduced breath support, making transcription more difficult. In this work, we benchmark acoustic foundation models on post-exercise speech under a unified evaluation protocol. We compare sequence-to-sequence models (Whisper and FunASR/Paraformer) and self-supervised encoders with CTC decoding (Wav2Vec2, HuBERT, and WavLM), under both off-the-shelf inference and post-exercise in-domain fine-tuning. Across the Static/Post-All benchmark, most models degrade on post-exercise speech, while FunASR shows the strongest baseline robustness at 14.57% WER and 8.21% CER on Post-All. Fine-tuning substantially improves several CTC-based models, whereas Whisper shows unstable adaptation. As an exploratory case study, we further stratify results by fluent and non-fluent speakers; although the non-fluent subset is small, it is consistently more challenging than the fluent subset. Overall, our findings show that post-exercise ASR robustness is strongly model-dependent, that in-domain adaptation can be highly effective but not uniformly stable, and that future post-exercise ASR studies should explicitly separate fluency-related effects from exercise-induced speech variation.

Automatic speech recognition (ASR) systems have achieved near-human accuracy on curated benchmarks, yet still fail in real-world voice agents under conditions that current evaluations do not systematically cover. Without diagnostic tools that isolate specific failure factors, practitioners cannot anticipate which conditions, in which languages, will cause what degree of degradation. We introduce WildASR, a multilingual (four-language) diagnostic benchmark sourced entirely from real human speech that factorizes ASR robustness along three axes: environmental degradation, demographic shift, and linguistic diversity. Evaluating seven widely used ASR systems, we find severe and uneven performance degradation, and model robustness does not transfer across languages or conditions. Critically, models often hallucinate plausible but unspoken content under partial or degraded inputs, creating concrete safety risks for downstream agent behavior. Our results demonstrate that targeted, factor-isolated evaluation is essential for understanding and improving ASR reliability in production systems. Besides the benchmark itself, we also present three analytical tools that practitioners can use to guide deployment decisions.

With the increasing deployment of automated and agentic systems, ensuring the adversarial robustness of automatic speech recognition (ASR) models has become highly relevant. We observe that changing the precision of an ASR model during inference reduces the likelihood of adversarial attacks to succeed. We take advantage of this fact to make models more robust simply by randomly sampling the precision during prediction. Moreover, this insight can be turned into an adversarial example detection strategy by implementing a simple Gaussian classifier that thresholds the differences between outputs of models run with different precision. To further enhance security boundaries, we combine the approach with an existing uncertainty-based defense mechanism, which forces adaptive adversaries to introduce highly perceptible noise to bypass detection. An experimental analysis across various ASR models, languages, and attack types demonstrates a significant increase in adversarial robustness, competitive detection capabilities, and resistance to adaptive threats.

We present Artoo, a learned acoustic communication system for robots that replaces hand-designed signal processing with end-to-end co-trained neural networks. Our system pairs a lightweight text-to-speech (TTS) transmitter (1.18M parameters) with a conformer-based automatic speech recognition (ASR) receiver (938K parameters), jointly optimized through a differentiable channel. Unlike human speech, robot-to-robot communication is paralinguistics-free: the system need not preserve timbre, prosody, or naturalness, only maximize decoding accuracy under channel distortion. Through a three-phase co-training curriculum, the TTS transmitter learns to produce distortion-robust acoustic encodings that surpass the baseline under noise, achieving 8.3% CER at 0 dB SNR. The entire system requires only 2.1M parameters (8.4 MB) and runs in under 13 ms end-to-end on a CPU, making it suitable for deployment on resource-constrained robotic platforms.

Recently, Automatic Speech Recognition (ASR) systems (e.g., Whisper) have achieved remarkable accuracy improvements but remain highly sensitive to real-world unseen data (data with large distribution shifts), including noisy environments and diverse accents. To address this issue, test-time adaptation (TTA) has shown great potential in improving the model adaptability at inference time without ground-truth labels, and existing TTA methods often rely on pseudo-labeling or entropy minimization. However, by treating model confidence as a learning signal, these methods may reinforce high-confidence errors, leading to confirmation bias that undermines adaptation. To overcome these limitations, we present ASR-TRA, a novel Test-time Reinforcement Adaptation framework inspired by causal intervention. More precisely, our method introduces a learnable decoder prompt and utilizes temperature-controlled stochastic decoding to generate diverse transcription candidates. These are scored by a reward model that measures audio-text semantic alignment, and the resulting feedback is used to update both model and prompt parameters via reinforcement learning. Comprehensive experiments on LibriSpeech with synthetic noise and L2 Arctic accented English datasets demonstrate that our method achieves higher accuracy while maintaining lower latency than existing TTA baselines. Ablation studies further confirm the effectiveness of combining audio and language-based rewards, highlighting our method's enhanced stability and interpretability. Overall, our approach provides a practical and robust solution for deploying ASR systems in challenging real-world conditions.

Audio-Visual Speech Recognition (AVSR) integrates acoustic and visual information to enhance robustness in adverse acoustic conditions. Recent advances in Large Language Models (LLMs) have yielded competitive automatic speech recognition performance and shown effectiveness for AVSR. However, prior approaches project audio and visual features independently or apply shallow fusion, limiting cross-modal alignment and complementary exchange while increasing the LLM's computational load. To address this, we propose AVUR-LLM, an LLM-based Audio-Visual Speech Recognition via Sparse Modality Alignment and Visual Unit-Guided Refinement. Experiments on LRS3 demonstrate state-of-the-art results for AVSR. Under additive-noise conditions at 0 dB SNR, it achieves 37% relative improvement over the baseline system.

We introduce Whisper-RIR-Mega, a benchmark dataset of paired clean and reverberant speech for evaluating automatic speech recognition (ASR) robustness to room acoustics. Each sample pairs a clean LibriSpeech utterance with the same utterance convolved with a real room impulse response from the RIR-Mega corpus, with stratified splits by reverberation time (RT60) and direct-to-reverberant ratio (DRR). We evaluate five Whisper models (tiny through large-v3) on 1600 test samples and report word error rate (WER) and character error rate (CER) under clean and reverberant conditions. Reverberation consistently degrades performance across all model sizes; the reverb penalty in WER ranges from 2.31 to 15.50 percentage points depending on the model. Whisper-large-v3 shows the smallest penalty; Whisper-tiny shows the largest. We release the dataset, evaluation code, and baseline results to support reproducible research on robust ASR.