Eliya Nachmani
I recently completed Ph.D. at Tel-Aviv University (advised by Prof. Lior Wolf) and was Researcher at Facebook AI Research (FAIR).
Research Area
Deep Learning
Audio & Speech
Signal Processing
Information theory & Coding theory
Publications
2024
E. Nachmani, A. Levkovitch, J. Salazar, C. Asawaroengchai, S. Mariooryad, RJ Skerry-Ryan, M. Ramanovich - LMs with a Voice: Spoken Language Modeling beyond Speech Tokens - ICLR 2024
S. Lutati, E. Nachmani, L. Wolf - Separate and Diffuse: Using a Pretrained Diffusion Model for Improving Source Separation - ICLR 2024
E. Nachmani, A. Levkovitch, Y. Ding, C. Asawaroengchai, H. Zen, M. Ramanovich - Translatotron 3: Speech to Speech Translation with Monolingual Data - ICASSP 2024
2023
S.Bar-David, I.Zimerman, E.Nachmani, L.Wolf - Decision S4: Efficient Sequence-Based RL via State Spaces Layers - ICLR
S.Sheynin, O.Ashual, A.Polyak, U.Singer, O.Gafni, E.Nachmani, Y.Taigman - KNN-Diffusion: Image Generation via Large-Scale Retrieval - ICLR
Y.Ger, E.Nachmani, L.Wolf, N.Shahar - Harnessing the Flexibility of Neural Networks to Predict Meaningful Theoretical Parameters in a Multi-Armed Bandit Task - Cosyne
2022
E. Nachmani, Y. Be'ery - Neural Decoding with Optimization of Node Activations - IEEE Communications Letters
A Levkovitch, E. Nachmani, L. Wolf - Zero-Shot Voice Conditioning for Denoising Diffusion TTS Models, Audio Samples - InterSpeech
S. Lutati, E. Nachmani, L. Wolf - SepIt: Approaching a Single Channel Speech Separation Bound - InterSpeech
O. Goren, E. Nachmani, L. Wolf - A-Muze-Net: Music Generation by Composing the Harmony based on the Generated Melody - MMM
K. Tan, B. Xu, A. Kumar, E. Nachmani, Y. Adi - SAGRNN: Self-Attentive Gated RNN for Binaural Speaker Separation with Interaural Cue Preservation - ICASSP
2021
T. Amit, E. Nachmani, T. Shaharbany, L. Wolf - SegDiff: Image Segmentation with Diffusion Probabilistic Models - Arxiv
E. Nachmani*, S. Dovrat* - Zero-Shot Translation using Diffusion Models - preprint
E. Nachmani*, R. San Roman*, L. Wolf - Non Gaussian Denoising Diffusion Models - Arxiv (*Equal contribution)
I. Zimerman*, E. Nachmani*, L. Wolf - Recovering AES Keys with a Deep Cold Boot Attack - ICML (*Equal contribution)
S. Dovrat*, E. Nachmani*, L. Wolf - Many-Speakers Single Channel Speech Separation with Optimal Permutation Training - INTERSPEECH (*Equal contribution)
E. Nachmani, L. Wolf - Autoregressive Belief Propagation for Decoding Block Codes - Arxiv
R. San-Roman*, E. Nachmani*, L. Wolf - Noise Estimation for Generative Diffusion Models - Arxiv
SE. Chazan, L. Wolf, E. Nachmani, Y. Adi - Single Channel Voice Separation for Unknown Number of Speakers under Reverberant and Noisy Settings - ICASSP
2020
K. Tan, B. Xu, A. Kumar, E. Nachmani, Y. Adi - SAGRNN: Self-Attentive Gated RNN for Binaural Speaker Separation with Interaural Cue Preservation, Audio Samples - IEEE Signal Processing Letters (IEEE SPL)
E. Nachmani, Y. Adi, L. Wolf Voice Separation with an Unknown Number of Multiple Speakers, Audio Samples, Github - ICML
E. Nachmani, L. Wolf A Gated Hypernet Decoder for Polar Codes - ICASSP
E. Nachmani, L. Wolf Molecule Property Prediction and Classification with Graph Hypernetworks - preprint
2019
E. Nachmani, L. Wolf Hyper-Graph-Network Decoders for Block Codes , Github - NeurIPS
E. Nachmani, L. Wolf Unsupervised Singing Voice Conversion - INTERSPEECH
E. Nachmani, L. Wolf Unsupervised Polyglot Text-to-Speech - ICASSP
2018
E. Nachmani, A. Polyak, Y. Taigman, L. Wolf Fitting New Speakers Based on a Short Untranscribed Sample, Talk - ICML
E. Nachmani, E. Marciano, L. Lugosch, WJ. Gross, D. Burshtein, Y. Be’ery Deep Learning Methods for Improved Decoding of Linear Codes. IEEE Journal of Selected Topics in Signal Processing
E. Nachmani, Y. Bachar, E. Marciano, D. Burshtein, Y. Be'ery Near Maximum Likelihood Decoding with Deep Learning - IZS
Y. Taigman, L. Wolf, A. Polyak, E. Nachmani VoiceLoop: Voice Fitting and Synthesis via a Phonological Loop - ICLR
2017
E. Nachmani, E. Marciano, D. Burshtein, Y. Be'ery RNN Decoding of Linear Block Codes - preprint
E. Nachmani, Y. Be'ery, D. Burshtein Learning to Decode Linear Codes Using Deep Learning - Allerton Conference on Communication, Control, and Computing
Teaching
Deep Learning - TA, 0368-4488-01, Fall semester 2020/2021
Deep Learning - TA, 0368-4488-01, Fall semester 2019/2020
Deep Learning - TA, 0368-4488-01, Fall semester 2018/2019
Reinforcement Learning - TA, 0368-3075-01, Spring semester 2017
Deep Learning - TA, 0368-4488-01, Fall semester 2016/2017
Media
Posters
ICML 2021
Neurips 2019
ICASSP 2019
ICML 2018
ICLR 2018
News
1.2023: “Decision S4: Efficient Sequence-Based RL via State Spaces Layers” accepted to ICLR 2023
1.2023: “KNN-Diffusion: Image Generation via Large-Scale Retrieval” accepted to ICLR 2023
5.2021: “Recovering AES Keys with a Deep Cold Boot Attack” accepted to ICML 2021
5.2020: “Voice Separation with an Unknown Number of Multiple Speakers” accepted to ICML 2020
9.2019: "Hyper-Graph-Network Decoders for Block Codes" accepted to NeurIPS 2019
5.2018: “Fitting New Speakers Based on a Short Untranscribed Sample” accepted to ICML 2018
1.2018: "VoiceLoop: Voice Fitting and Synthesis via a Phonological Loop" accepted to ICLR 2018
Contact
