Denis Gudovskiy

Denis Gudovskiy

Senior Researcher at Panasonic AI Lab

About Me

Hi, my name is Denis and I work at Panasonic AI Lab in Mountain View office. My research projects belong to general Machine Learning area and, more specifically, they are focused on various aspects of recent Deep Learning methods for Computer Vision applications (see a short bio below). This is my personal page with recent publications, talks, and CV updates. In my free time, I enjoy riding motorcycles and mountaineering in Sierra Nevada.

News

  • May 2024: In collaboration with UC Berkeley, our paper called “Split-Ensemble: Efficient OOD-aware Ensemble via Task and Model Splitting” has been accepted to ICML’24! Preprint is available on ArXiv.
  • May 2024: Our paper called “ContextFlow++: Generalist-Specialist Flow-based Generative Models with Mixed-Variable Context Encoding” has been accepted to UAI’24! Preprint is available on ArXiv.
  • February 2024: In collaboration with UC Berkeley, our paper called “Efficient Deweather Mixture-of-Experts with Uncertainty-aware Feature-wise Linear Modulation” has been accepted to AAAI’24! Preprint is available on ArXiv.
  • May 2023: Our paper called “Concurrent Misclassification and Out-of-Distribution Detection for Semantic Segmentation via Energy-Based Normalizing Flow” has been accepted to UAI’23! Preprint is available on ArXiv.
  • September 2022: In collaboration with UC Berkeley, our paper called “MTTrans: Cross-Domain Object Detection with Mean-Teacher Transformer” has been accepted to ECCV’22! Preprint is available on ArXiv.
  • October 2021: Our paper called “Contrastive Neural Processes for Self-Supervised Learning” has been accepted to ACML’21 (oral)! Preprint is available on ArXiv.
  • August 2021: Our paper called “CFLOW-AD: Real-Time Unsupervised Anomaly Detection with Localization via Conditional Normalizing Flows” has been accepted to WACV’22! Preprint is available on ArXiv.
  • March 2021: Our paper called “AutoDO: Robust AutoAugment for Biased Data with Label Noise via Scalable Probabilistic Implicit Differentiation” has been accepted to CVPR’21! Preprint is available on ArXiv.
  • March 2020: Our paper called “Deep Active Learning for Biased Datasets via Fisher Kernel Self-Supervision” has been accepted to CVPR’20! Preprint is available on ArXiv.

Bio

Denis Gudovskiy is a senior researcher at Panasonic AI lab in Mountain View. He specializes in deep learning-based algorithms for AI applications. His portfolio of research projects includes optimization of deep neural networks for edge AI devices, explainable AI tools, and automatic dataset management for computer vision applications. Before joining Panasonic in 2016, Denis held research and engineering positions in Intel, Olympus and Huawei wireless divisions. Denis received his M.Sc. in Computer Engineering from the University of Texas, Austin in 2008.

In one of such projects he proposed and implemented a method of hardware-efficient neural network quantization and compression techniques for autonomous vehicles with stringent power and performance requirements. In the most recent work, Denis has demonstrated how data annotation costs could be significantly reduced with a use of advanced algorithms. His papers and demos are presented and published in top-tier machine learning and computer vision conferences such as NeurIPS, CVPR, ECCV, ICLR and ICASSP.

Denis sees corporate research as an important layer between moonshot academia projects and clearly-defined product development roadmaps in business units. His goal is to find and promote viable academia-grade opportunities at Panasonic within the exponentially growing landscape of AI applications.