Intel Showcases Deep Learning Research and Neuroscience Workshop at ICLR 2021
Intel researchers and collaborators present eight papers and an all-day workshop exploring the intersection of neuroscience and artificial intelligence at the virtual ICLR 2021 conference.
Intel researchers and their collaborators present eight papers alongside an all-day workshop at the virtual International Conference on Learning Representations (ICLR) from May 3-7, 2021. The conference serves as a major venue for showcasing cutting-edge research on all aspects of deep learning as it applies to artificial intelligence, statistics, data science, and critical application areas like computer vision, natural language processing, and robotics.
On May 7th, Intel hosts an intensive workshop titled "How Can Findings About The Brain Improve AI Systems?" to explore how neuroscientific insights lead to better AI. Participants investigate which areas of AI benefit most from neuroscience, address current bottlenecks in integrating biological data with machine learning, and discuss the advantages and limits of using this data to train systems for complex tasks like natural language processing and vision.
The submitted Intel research papers cover a wide range of deep learning topics that highlight the company's broad expertise in the field. One featured paper introduces an original energy-based solution for semi-supervised visual anomaly detection and localization, which identifies arbitrary anomalies in test data using only anomaly-free training examples, while another paper applies massively parallel Monte-Carlo Tree Search to the complex field of molecular design.