Kunal Rathore

Kunal Rathore

PhD student in Environmental Science & AI

Posts by Collection

portfolio

Colorful Corvallis

Published:

Blooming Beauties: colored memories I can not ever leave behind!!

projects

XAI-Via-Correspondence

Published:

This page provides details about my research work in Explainable AI at a brief level.

publications

A Simple Method of Solution For Multi-label Feature Selection

Published in IEEE International Conference on Electrical, Computer and Communication Technologies, 2019

The study proposes a two-step algorithm for multi-label classification (MLC). It first decomposes the output label space into lower dimensions, then applies feature selection in the reduced space. This approach efficiently handles high-dimensional datasets and reduces computational load.

Recommended citation: Valadi, Jayaraman K., Prasad T. Ovhal, and Kunal J. Rathore. "A simple method of solution for multi-label feature selection." 2019 IEEE International Conference on Electrical, Computer and Communication Technologies (ICECCT). IEEE, 2019. https://ieeexplore.ieee.org/document/8869493

RA2Vec: Distributed representation of protein sequences with reduced alphabet embeddings

Published in Proceedings of the 11th ACM international conference on bioinformatics, computational biology and health informatics., 2020

This study introduces RA2Vec, a new method for protein function identification using reduced amino acid alphabets. It maps Swiss-Prot sequences to a reduced form based on hydropathy and conformational similarity. The method uses a skip-gram approach to create embedding vectors for each set. These vectors are then used as input to Support Vector Machines classifiers. The vectors are further refined using recursive Feature Elimination to maximize accuracy. The results show that certain combinations of these new representations can significantly improve performance.

Recommended citation: Wijesekara, Rajitha Yasas, et al. "RA2Vec: Distributed representation of protein sequences with reduced alphabet embeddings: RA2Vec: distributed representation." Proceedings of the 11th ACM international conference on bioinformatics, computational biology and health informatics. 2020. https://dl.acm.org/doi/abs/10.1145/3388440.3414925

Generating Part-based Global Explnations via Correspondence

Published in IJCAI 2024 Workshop on Explainable Artificial Intelligence (XAI), 2024

Deep learning models are notoriously opaque. Existing explanation methods often focus on localized visual explanations for individual images. Concept-based explanations, while offering global insights, require extensive annotations, incurring significant labeling cost. We propose an approach that leverages user-defined part labels from a limited set of images and efficiently transfers them to a larger dataset. This enables the generation of global symbolic explanations by aggregating part-based local explanations, ultimately providing human-understandable explanations for model decisions on a large scale.

Recommended citation: Kunal Rathore, (2024). "Generating Part-based Global Explnations via Correspondence." IJCAI 1. 1(1). https://drive.google.com/file/d/1_1LySYzjUUow22PV4M9FScRstdFIzk89/view

talks

Poster: Explanations via Minimal regions and Semantic Correspondence

Published:

The objective of this poster presentation was to articulate our research endeavors in the realm of Explainable Artificial Intelligence to industry professionals, and to illustrate potential applications thereof. The central premise was to underscore the importance of generating human-comprehensible and interpretable explanations through sub-symbolic representations of images, such as components of an object within an image, which serve as inputs to a Deep Neural Network (DNN) model.

Poster: Discriminative Global Explanations via Part Attributes

Published:

This poster presentation served as a platform for me to articulate our research concepts and advancements in the field of Explainable Artificial Intelligence to industry professionals. Subsequently, I elaborated on the discriminative global explanations and the methodology we are employing to generate such explanations, along with their potential benefits to the industry.

XAI@IJCAI: Generating Part-Based Global Explanations Via Correspondence

Published:

We proposed a methodology for generating global explanations using correspondence techniques. The focus was on delivering human-comprehensible verbal explanations, articulated as part-based expressions, derived from the input image to the Deep Neural Network classifier.

teaching

Teaching experience 1

Undergraduate course, University 1, Department, 2014

This is a description of a teaching experience. You can use markdown like any other post.

Teaching experience 2

Workshop, University 1, Department, 2015

This is a description of a teaching experience. You can use markdown like any other post.