What makes a Model Compositional? A Neuro-Symbolic Theoretical View
- Parikshit Ram
- Tim Klinger
- et al.
- 2024
- IJCAI 2024
Parikshit Ram is a Principal Research Staff Member in IBM Research, with broad research interests in similarity search, efficient all-pairs algorithms, density estimation, computational geometry, and automated machine learning and data science among other things. He currently conducts basic mathematical and applied computational research on topics pertinent to optimization as well as various aspects of generalization and learning with less data. Prior to joining IBM Research, he was a Senior Research Staff Member at Skytree, a machine learning company focused on providing high performance machine learning tools for large scale modeling and data analysis, which was subsequently acquired by Infosys. Parikshit received his Doctorate in machine learning at Georgia Institute of Technology and a Bachelors and a Masters in Mathematics and Computing from the Indian Institute of Technology, Kharagpur. He has been in the program committee and senior program committee of top conferences and recognized as a top reviewer at ICML and NeurIPS multiple times. More details on his research can be found at https://rithram.github.io.