Taras Khakhulin

I am research engineer at at Synthesia , where we’re working on the future of media 🚀. My current focus is on developing generative models based on diffusion processes for video and motion. Before I was a Ph.D. student at Skoltech advised by Victor Lempitsky, and worked as a research engineer at Samsung AI Center for several years, where we pushed the boundaries of view synthesis. I try to synthesize things that people currently cannot. During my studies, I have contributed to 3D representations, image synthesis, and human avatars.

Prior to that, I investigated the application of reinforcement learning for discrete optimization under the advising of Ivan Oseledets. Earlier, I was a part of the DeepPavlov project, where we studied word representations for text with missplellings using context-aware language modeling.

Email  /  CV  /  Google Scholar  /  Twitter  /  LinkedIn  /  Github

profile photo
Research
HumanRF: High-Fidelity Neural Radiance Fields for Humans in Motion
Mustafa Işık, Martin Rünz, Markos Georgopoulos, Taras Khakhulin, Jonathan Starck, Lourdes Agapito, Matthias Nießner
ACM ToG, SIGGRAPH, 2023
project page / arXiv / code

Introduced high-resolution dataset of human performances to speed up development in avatars. Also, we demonstrated ability to reconstruct large scenes with temporal factorization for 4D NeRFs.

Self-improving Multiplane-to-layer Images for Novel View Synthesis
Pavel Solovev*, Taras Khakhulin*, Denis Korzhenkov*
WACV, 2023
project page / arXiv / immersive results / code / bibtex

Create the most efficient renderable representtion for novel view synthesis from an arbitrary number of images (more than one). We learn this system end-to-end on the dataset of monocular videos. Extremely fast and compact way to present scene.

Realistic One-shot Mesh-based Head Avatars
Taras Khakhulin, Vanessa Sklyarova, Victor Lempitsky, Egor Zakharov
ECCV, 2022
project page / arXiv / code / bibtex

Create an animatable avatar just from a single image with coarse hair mesh and neural rendering. We learn head geometry and rendering together with supreme quality in a cross-person reenactment.

MegaPortraits: One-shot Megapixel Neural Head Avatars
Nikita Drobyshev, Jenya Chelishev, Taras Khakhulin, Aleksei Ivakhnenko, Victor Lempitsky, Egor Zakharov
ACMM, 2022
project page / arXiv / bibtex

One-shot high-resolution avatars with latent poses for cross-reenactment.

Stereo Magnification with Multi-Layer Images
Taras Khakhulin, Denis Korzhenkov, Pavel Solovev, Gleb Sterkin, Andrei-Timotei Ardelean, Victor Lempitsky
CVPR, 2022
project page / arXiv / dataset / blog / bibtex

The scene can be represented as a set of semi-transparent mesh layers from just stereo pair without loss of the quality. This representation allows effortless estimation and fast rendering. Additionaly, we published a dataset with occluded region - SWORD - for novel view synthesis.

Image Generators with Conditionally-Independent Pixel Synthesis
Ivan Anokhin, Kiril Demochkin, Taras Khakhulin, Gleb Sterkin,
Victor Lempitsky, Denis Korzhenkov
CVPR, 2021 (Oral Presentation)
arXiv / code / samples / blog / bibtex

Our generator produce images without any spatial convolutions. Each pixel synthesized separetly conditioned on noise vector. We investigate properties of such generator and propose several applications (e.g. super-res, foveated rendering).

Learning Elimination Ordering for Tree Decomposition Problem
Taras Khakhulin Roman Schutski, Ivan Oseledets
Learning Meets Combinatorial Algorithms, NeurIPS Workshop , 2020
presentation / paper / arXiv / bibtex

We propose a learning heuristic with RL for real-world combinatorial problem on graphs. Suprisingly, we can easily estimate universal policy which can be scaled across different graphs.

High-Resolution Daytime Translation Without Domain Labels
Ivan Anokhin*, Pavel Solovev*, Denis Korzhenkov*, Alexey Kharlamov*
Taras Khakhulin Alexey Silvestrov, Sergey Nikolenko, Victor Lempitsky, Gleb Sterkin
CVPR, 2020 (Oral Presentation)
video / project / arXiv / code / bibtex

Our generator produce images without any spatial convolutions. Each pixel synthesized separetly conditioned on noise vector.

td Simple heuristics for efficient parallel tensor contraction and quantum circuit simulation
Roman Schutski, Dmitry Kolmakov, Taras Khakhulin, Ivan Oseledets
Physical Review A, 2020
arXiv / bibtex

The heuristic approach based on tree decomposition to relax the contraction of tensor networks using probabilistic graphical models and applied for random quantum circuits.

rove Robust word vectors: context-informed embeddings for noisy texts
Valentin Malykh, Varvara Logacheva, Taras Khakhulin
Workshop on Noisy User-generated Text at EMNLP, 2018
paper / bibtex

We suggest a new language-independent architecture of robust word vectors (RoVe). It is designed to alleviate the issue of typos, which are common in almost any user-generated content, and hinder automatic text processing. Our model allows it to deal with unseen word forms in morphologically rich languages before contextualized language models.

deeppavlov Deeppavlov: Open-source library for dialogue systems
Mikhail Burtsev, Alexander Seliverstov, Rafael Airapetyan, Mikhail Arkhipov, Dilyara Baymurzina, Nickolay Bushkov, Olga Gureenkova, Taras Khakhulin , Yuri Kuratov, Denis Kuznetsov, Alexey Litinsky, Varvara Logacheva, Alexey Lymar, Valentin Malykh, Maxim Petrov, Vadim Polulyakh, Leonid Pugachev, Alexey Sorokin, Maria Vikhreva, Marat Zaynutdinov
ACL, System Demonstrations , 2018
paper / code / TF blog / bibtex

Open-source library for end2end conversational NLP.

The webpage template was borrowed from Jon Barron.