REVIEW 24 cited by
FLEURS: Few-shot Learning Evaluation of Universal Representations of Speech
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
Signed reviews
read the original abstract
We introduce FLEURS, the Few-shot Learning Evaluation of Universal Representations of Speech benchmark. FLEURS is an n-way parallel speech dataset in 102 languages built on top of the machine translation FLoRes-101 benchmark, with approximately 12 hours of speech supervision per language. FLEURS can be used for a variety of speech tasks, including Automatic Speech Recognition (ASR), Speech Language Identification (Speech LangID), Translation and Retrieval. In this paper, we provide baselines for the tasks based on multilingual pre-trained models like mSLAM. The goal of FLEURS is to enable speech technology in more languages and catalyze research in low-resource speech understanding.
Forward citations
Cited by 24 Pith papers
-
Multilingual Phonological Feature Recognition with Self-Supervised Speech Models
PhonoQ-2.0 directly predicts structured phonological features from self-supervised models with a gating mechanism, outperforming phoneme baselines by 8+ macro-F1 points on average across in-domain, out-of-domain, and ...
-
myMediWhisper: Construction of Burmese Medical Speech Corpus and Whisper Fine-Tuning for Clinical Dialogue ASR
A new Burmese medical speech corpus is released and Whisper fine-tuning (full and LoRA) is benchmarked, with the best model reaching 23.44% WER.
-
Compress the Cache, Not the Speech Embedding: KV Compression for Efficient Speech LLMs
Learned pooling of speech KV caches from an intermediate LLM layer compresses speech to text-level length while matching or exceeding the uncompressed baseline on ASR and entity recognition, with 1.49–2× decoding speedup.
-
Listen, Think, Transcribe: Continuous Latent Test-Time Scaling for ASR
Two small modules enable continuous latent test-time refinement on a frozen ASR backbone, cutting error on hard speech under a 500-utterance regime where fine-tuning, LoRA and prompt tuning all regress.
-
Improving French Synthetic Speech Quality via SSML Prosody Control
Two fine-tuned LLMs predict SSML prosody tags that raise French TTS naturalness from a 3.20 to 3.87 MOS.
-
Improving Language and Modality Transfer in Translation by Character-level Modeling
A character-level encoder distilled from SONAR embeddings improves cross-lingual transfer, and a pretrained adapter connects MMS speech recognition to it for competitive zero-shot speech translation.
-
TEDI: Trustworthy and Ethical Dataset Indicators to Analyze and Compare Dataset Documentation
A new 143-indicator rubric applied to 114 human-voice datasets shows that documentation of consent, privacy, and harmful content is rare, and that scraping yields scale at the cost of documented ethical practices.
-
Fleurs-SLU: A Massively Multilingual Benchmark for Spoken Language Understanding
Fleurs-SLU provides spoken topic classification in 102 languages and spoken multiple-choice QA in 92 languages, with comparisons of speech encoders, cascaded ASR+LLM systems, and speech-LLMs.
-
2M-BELEBELE: Highly Multilingual Speech and American Sign Language Comprehension Dataset
2M-BELEBELE is a new multilingual speech and ASL comprehension benchmark built from BELEBELE and FLEURS, with human recordings for 74 spoken languages and ASL video with glosses.
-
Raon-Speech Technical Report
A 9B SpeechLM trained on 1.38M hours of English/Korean data plus a full-duplex chat extension trained on 119K hours of time-aligned dialogue outperform same-size audio models on speech tasks and FDB turn-taking metrics.
-
A Semi-spontaneous Dutch Speech Dataset for Speech Enhancement and Speech Recognition
On real noisy Dutch semi-spontaneous speech, five of eight SOTA ASR models reach WER under 22%, yet five single-channel SE methods fail to improve and often degrade ASR.
-
ParaASR: Multi-Token Prediction for Fast and Long-Context LLM-Based Speech Recognition
A 4B-parameter LLM ASR system with five multi-token-prediction branches reports 2.97% CER Chinese, 3.68% WER English, 3.70% long-form WER, and a 0.0053 real-time factor, but the acceptance-rate calculation and ablatio...
-
MEUSLI: a Multilingual Projector for LLM-based ASR and Beyond
MEUSLI, a released family of linear projectors, enables open-source LLM-based ASR across 28 European languages and supports few-hour adaptation to unseen languages and extra speech tasks.
-
X-Translator: A Real-Time Multilingual Speaker-Aware Speech-to-Speech Translation System
An open, modular cascaded system (streaming ASR + MT + prompt-conditioned TTS) preserves speaker identity in long-form multi-speaker translation, at higher latency and slightly lower translation quality than proprietary APIs.
-
FastSLM: Hierarchical Temporal Abstraction for Efficient Long-Form Speech Adaptation
A hierarchical Q-Former compresses speech to about 1.67 tokens/sec, enabling hour-long audio processing with near-linear memory scaling and competitive benchmark scores.
-
RoboEgo System Card: An Omnimodal Model with Native Full Duplexity
RoboEgo combines a full-duplex audio/text backbone with vision and action heads to achieve an 80 ms theoretical response granularity, reporting competitive quality and better responsiveness than Qwen2.5-Omni in a smal...
-
Lillama: Large Language Models Compression via Low-Rank Feature Distillation
Lillama compresses LLMs by SVD-initialized low-rank layers trained with a local Teacher plus Student activation distillation loss, achieving 20-40% parameter reduction with only 13 million calibration tokens.
-
Typhoon 2: A Family of Open Text and Multimodal Thai Large Language Models
Typhoon 2 improves Thai LLM performance through continual pre-training on curated Thai data and post-training, releasing text, vision, audio, and safety models.
-
Technical report: Impact of Duration Prediction on Speaker-specific TTS for Indian Languages
In a five-language zero-shot TTS study, no single duration prediction strategy dominates: speaker-prompted durations help some languages, infilling durations help others, and results vary by metric.
-
A2TTS: TTS for Low Resource Indian Languages
A2TTS adds a reference-audio cross-attention duration predictor to a Grad-TTS and UnitSpeech style diffusion TTS, improving speaker similarity scores in seven Indian languages.
-
HumanOmni: A Large Vision-Speech Language Model for Human-Centric Video Understanding
A human-centric vision-speech language model with three instruction-weighted visual branches and audio input reports strong emotion, facial expression, and action results, but its evaluation is under-specified and not...
-
Deep Learning-Based Feature Fusion for Emotion Analysis and Suicide Risk Differentiation in Chinese Psychological Support Hotlines
A speech emotion recognition model fusing Wav2Vec2, pitch, and MFCC features with dual attention achieved 79.13% F1 on hotline negative emotion classification and 73.30% on a Vietnamese benchmark, but suicide-group em...
-
High-precision medical speech recognition through synthetic data and semantic correction: UNITED-MEDASR
A synthetic-data pipeline for medical ASR reports sub-1% WER on standard benchmarks, but the reported numbers are internally inconsistent and not reproducible from the paper.
-
Whisper Turns Stronger: Augmenting Wav2Vec 2.0 for Superior ASR in Low-Resource Languages
Fine-tuning Wav2Vec2-xlsr-53 on Common Voice audio augmented with pitch shift, Gaussian noise, and band-stop filtering lowers WER and CER in Arabic, Russian, and Portuguese, though the Whisper comparison is overstated.
Discussion (0). Continue with ORCID to comment.