Posts by Collection

Tran, Ba-Hien; Milios, Dimitrios; Rossi, Simone; Filippone, Maurizio. Functional Priors for Bayesian Neural Networks through Wasserstein Distance Minimization to Gaussian Processes. 3rd Symposium on Advances in Approximate Bayesian Inference, 2021.

Model Selection for Bayesian Autoencoders

Published in NeurIPS, 2021

Tran, Ba-Hien; Rossi, Simone; Milios, Dimitrios; Michiardi, Pietro; V. Bonilla, Edwin; Filippone, Maurizio. Model Selection for Bayesian Autoencoders. In Advances in Neural Information Processing Systems, 2021.

All You Need is a Good Functional Prior for Bayesian Deep Learning

Published in JMLR, 2022

Tran, Ba-Hien; Rossi, Simone; Milios, Dimitrios; Filippone, Maurizio. All You Need is a Good Functional Prior for Bayesian Deep Learning. Journal of Machine Learning Research, 2022.

Fully Bayesian Autoencoders with Latent Sparse Gaussian Processes

Published in ICML, 2023

Tran, Ba-Hien; Shahbaba, Babak; Mandt, Stephan; Filippone, Maurizio. Fully Bayesian Autoencoders with Latent Sparse Gaussian Processes. International Conference on Machine Learning, 2023.

Improving Training of Likelihood-based Generative Models with Gaussian Homotopy

Published in ICML Workshop on Structured Probabilistic Inference and Generative Modeling, 2023

Tran, Ba-Hien; Franzese, Giulio; Michiardi, Pietro; Filippone, Maurizio. Improving Training of Likelihood-based Generative Models with Gaussian Homotopy. ICML Workshop on Structured Probabilistic Inference and Generative Modeling, 2023.

One-Line-of-Code Data Mollification Improves Optimization of Likelihood-based Generative Models

Published in NeurIPS, 2023

Tran, Ba-Hien; Franzese, Giulio; Michiardi, Pietro; Filippone, Maurizio. One-Line-of-Code Data Mollification Improves Optimization of Likelihood-based Generative Models. In Advances in Neural Information Processing Systems, 2023.

Spatial Bayesian Neural Networks

Published in Spatial Statistics, 2024

Zammit-Mangion, Andrew; Kaminski, D. Michael; Tran, Ba-Hien; Filippone, Maurizio; Cressie, Noel. Spatial Bayesian Neural Networks. In Spatial Statistics, Volume 60, 2024.

Boolean Logic for Low-Energy Deep Learning

Published in ICML Workshop on Structured Probabilistic Inference and Generative Modeling, 2024

Nguyen, Van Minh; Ocampo, Cristian; Askri, Aymen; Leconte, Louis; Tran, Ba-Hien. Boolean Logic for Low-Energy Deep Learning. ICML Workshop on Advancing Neural Network Training: Computational Efficiency, Scalability, and Resource Optimization, 2024.

teaching

Teaching experience 1

Undergraduate course, University 1, Department, 2014

This is a description of a teaching experience. You can use markdown like any other post.

Teaching experience 2

Workshop, University 1, Department, 2015

This is a description of a teaching experience. You can use markdown like any other post.

Ba-Hien TRAN

Posts by Collection

news

I have joined EURECOM as a PhD student in Machine Learning.

I have defended successfully my master thesis at Telecom Paris and EURECOM.

Check out our new paper: ‘All You Need is a Good Functional Prior for Bayesian Deep Learning’!

The paper ‘Functional Priors for Bayesian Neural Networks through Wasserstein Distance Minimization to Gaussian Processes’ has been accepted at the 3rd Symposium on Advances in Approximate Bayesian Inference 2021!

Invited talk at the Data Centric Engineering group, Alan Turing Institute, UK (virtual): ‘Functional priors for Bayesian neural networks’.

Check out our new paper: ‘Model Selection for Bayesian Autoencoders’!

Invited to serve as a reviewer for AISTATS.

Our paper ‘Model Selection for Bayesian Autoencoders’ has been accepted at NeurIPS 2021!

Accepted for a contributed talk at ISBA 2022 World Meeting on ‘Functional Priors for Bayesian Deep Learning’.

Invited talk at the SIAM Conference on Imaging Science (IS22), organized by Society for Industrial and Applied Mathematics.

Received an ISBA Travel Award for Young Researchers.

Our paper ‘All You Need is a Good Functional Prior for Bayesian Deep Learning’ has been accepted by JMLR!

Accepted for a contributed talk at Joint Statistical Meetings (JSM) 2022 on ‘Functional Priors for Bayesian Deep Learning’.

I’ve had a California trip during April 26th ~ May 8th to visit and give talks at Ermon’s lab (Stanford) and Mandt’s lab (UCI).

I’m visiting the groups of Professors Stephan Mandt and Babak Shahbaba at UCI for three months.

I will be attending the NeurIPS conference and presenting our JMLR paper.

Check out our new paper: ‘Fully Bayesian Autoencoders with Latent Sparse Gaussian Processes’!

Our paper ‘Fully Bayesian Autoencoders with Latent Sparse Gaussian Processes’ has been accepted at ICML 2023!

Check out our new paper: ‘One-Line-of-Code Data Mollification Improves Optimization of Likelihood-based Generative Models’!

Our paper ‘Improving Training of Likelihood-based Generative Models with Gaussian Homotopy’ has been accepted at ICML Workshop on Structured Probabilistic Inference and Generative Modeling 2023!

I will be attending and presenting a poster at the workshop ‘Generative models and uncertainty quantification’, GenU 2023.

Our paper ‘One-Line-of-Code Data Mollification Improves Optimization of Likelihood-based Generative Models’ has been accepted at NeurIPS 2023!

I have defended successfully my PhD thesis! The members of the jury include Chris Oates, Mark van der Wilk, Marco Lorenzi, Serena Villata, Pietro Michiardi, and Maurizio Filippone.

Our paper ‘Spatial Bayesian Neural Networks’ has been accepted in the Spatial Statistics journal!

Check out our new paper: ‘BOLD: Boolean Logic Deep Learning’!

Check out our new paper: ‘Robust Classification by Coupling Data Mollification with Label Smoothing’!

Our paper ‘Boolean Logic for Low-Energy Deep Learning’ has been accepted at ICML Workshop on Advancing Neural Network Training: Computational Efficiency, Scalability, and Resource Optimization 2024!

Our paper ‘BOLD: Boolean Logic Deep Learning’ has been accepted at NeurIPS 2024!

Our paper ‘Robust Classification by Coupling Data Mollification with Label Smoothing’ has been accepted at AISTATS 2025!

Check out our new paper: ‘Optimizing Data Augmentation through Bayesian Model Selection’!

Check out our new paper: ‘Highly Efficient and Effective LLMs with Multi-Boolean Architectures’!

Our paper ‘Ultra-Efficient and Effective Large Language Models with Multi-Boolean Architectures’ has been accepted at ICML Workshop on on Efficient Systems for Foundation Models and workshop on Machine Learning for Wireless Communication and Networks 2025!

Check out our new paper: ‘Universal Adaptive Environment Discovery’!

Our papers ‘Highly Efficient and Effective LLMs with Multi-Boolean Architectures’ and ‘Optimizing Data Augmentation through Bayesian Model Selection’ have been accepted at ICLR 2026!

portfolio

publications

talks

teaching