Tianlin Liu
Tianlin Liu
PhD candidate, University of Basel

I am a PhD candidate in machine learning at the University of Basel in Switzerland. I work with Prof. Ivan Dokmanić.

In the fall of 2023, I interned at Google DeepMind in Paris. In the summer of 2022, I interned at Google Brain in Zurich.

I received BSc (2016) and MSc (2019) both from Jacobs University Bremen. During my time there, I completed my master’s thesis under the guidance of Prof. Herbert Jaeger.

I welcome your feedback, which you can give anonymously here.


  • May 2024: Our paper on decoding-time realignment of language models was accepted by ICML 2024.

  • April 2024: Our paper on evaluating routers in vision mixture-of-experts was accepted by TMLR 2024.

  • Feb 2024: Our paper on benchmarking wave-propagation PDE solvers was accepted by TMLR, 2024.

  • Oct 2023: Check out our blog post for an in-depth look at OpenAI’s first RLHF paper, featuring results replicated in PyTorch and Jax. It has been selected for a spotlight in the ICLR-2024 blogpost track.

  • Jan 2023: Our paper on sparsity-constrained optimal transport was accepted by ICLR 2023 as a spotlight presentation.

  • May 2022: Our paper on multiscale convolutional dictionary learning was accepted by the IEEE Transactions on Computational Imaging.

  • Jan 2022: Our paper on universal approximation under constraints was by ICLR 2022 and selected for a spotlight presentation.

  • June 2020: Our paper on training sparse neural networks was accepted by ICML 2020.

  • Nov 2019: Our paper on word vector denoising was accepted by AAAI 2020.

  • March 2019: Our paper on training spiking neural networks for neuromorphic hardware won a best paper finalist award at IEEE NER-2019.