Get our free extension to see links to code for papers anywhere online!

Add to Chrome

Add to Firefox

Picture for Piotr Zelasko

Piotr Zelasko

Predicting Multi-Codebook Vector Quantization Indexes for Knowledge Distillation

Oct 31, 2022
Liyong Guo, Xiaoyu Yang, Quandong Wang, Yuxiang Kong, Zengwei Yao, Fan Cui, Fangjun Kuang, Wei Kang, Long Lin, Mingshuang Luo, Piotr Zelasko, Daniel Povey

Figure 1 for Predicting Multi-Codebook Vector Quantization Indexes for Knowledge Distillation

Figure 2 for Predicting Multi-Codebook Vector Quantization Indexes for Knowledge Distillation

Figure 3 for Predicting Multi-Codebook Vector Quantization Indexes for Knowledge Distillation

Figure 4 for Predicting Multi-Codebook Vector Quantization Indexes for Knowledge Distillation

* Submitted to ICASSP 2022

Access Paper or Ask Questions

Defense against Adversarial Attacks on Hybrid Speech Recognition using Joint Adversarial Fine-tuning with Denoiser

Apr 08, 2022
Sonal Joshi, Saurabh Kataria, Yiwen Shao, Piotr Zelasko, Jesus Villalba, Sanjeev Khudanpur, Najim Dehak

Figure 1 for Defense against Adversarial Attacks on Hybrid Speech Recognition using Joint Adversarial Fine-tuning with Denoiser

Figure 2 for Defense against Adversarial Attacks on Hybrid Speech Recognition using Joint Adversarial Fine-tuning with Denoiser

Figure 3 for Defense against Adversarial Attacks on Hybrid Speech Recognition using Joint Adversarial Fine-tuning with Denoiser

Figure 4 for Defense against Adversarial Attacks on Hybrid Speech Recognition using Joint Adversarial Fine-tuning with Denoiser

* Submitted to Interspeech 2022

Access Paper or Ask Questions

Earnings-21: A Practical Benchmark for ASR in the Wild

Apr 28, 2021
Miguel Del Rio, Natalie Delworth, Ryan Westerman, Michelle Huang, Nishchal Bhandari, Joseph Palakapilly, Quinten McNamara, Joshua Dong, Piotr Zelasko, Miguel Jette

Figure 1 for Earnings-21: A Practical Benchmark for ASR in the Wild

Figure 2 for Earnings-21: A Practical Benchmark for ASR in the Wild

Figure 3 for Earnings-21: A Practical Benchmark for ASR in the Wild

Figure 4 for Earnings-21: A Practical Benchmark for ASR in the Wild

* submitted to INTERSPEECH 2021 Update April 28th, 2021: We found and resolved an issue in our experimental evaluation that scored the LibriSpeech model at ~20% worse relative WER than the actual WER. The updated results do not affect our conclusions

Access Paper or Ask Questions