Deep Shankar Pandey

I’m an Applied Scientist at Amazon, where I work on search and recommendation systems that personalize content discovery for Fire TV customers.

My research is in trustworthy AI: developing models that can learn from limited data and give meaningful estimates of their uncertainty. During my Ph.D., I worked on evidential deep learning, Bayesian methods, and uncertainty-aware meta-learning, with an emphasis on efficient, robust, and well-calibrated few-shot learning.

I’m also interested in large language models and, more broadly, how to make AI systems reliable, adaptable, and useful in the real world.

I earned my Ph.D. in Computing and Information Sciences at Rochester Institute of Technology, where I worked with Prof. Qi Yu. I completed my undergraduate degree in Electronics and Communication Engineering at the Institute of Engineering, Pulchowk Campus in Nepal.

Publications Research & work Google Scholar

News

Jun 5, 2026	Our paper on generalized regularized evidential deep learning appeared in the June 2026 issue of IEEE TPAMI.
Jan 16, 2025	I defended my Ph.D. dissertation, Uncertainty-Aware Meta-Learning for Learning from Limited Data, at RIT.
Dec 10, 2024	Our work on Bayesian parameter-efficient fine-tuning of vision foundation models appeared at NeurIPS 2024.
Aug 19, 2023	I have successfully completed summer internship as Deep Learning for Image and Video Processing Intern at InterDigital Communications.

Selected Publications

TPAMI 2026

Generalized Regularized Evidential Deep Learning Models: Theory and Comprehensive Evaluation

Deep Shankar Pandey, Hyomin Choi, and Qi Yu

IEEE Transactions on Pattern Analysis and Machine Intelligence 2026
NeurIPS 2024

Be Confident in What You Know: Bayesian Parameter Efficient Fine-Tuning of Vision Foundation Models

Deep Shankar Pandey, Spandan Pyakurel, and Qi Yu

In Advances in Neural Information Processing Systems 2024
ICML 2023

Learn to Accumulate Evidence from All Training Samples: Theory and Practice

Deep Pandey, and Qi Yu

In Proceedings of the 40th International Conference on Machine Learning 2023

Abs

Evidential deep learning, built upon belief theory and subjective logic, offers a principled and computationally efficient way to turn a deterministic neural network uncertainty-aware. The resultant evidential models can quantify fine-grained uncertainty using the learned evidence. To ensure theoretically sound evidential models, the evidence needs to be non-negative, which requires special activation functions for model training and inference. This constraint often leads to inferior predictive performance compared to standard softmax models, making it challenging to extend them to many large-scale datasets. To unveil the real cause of this undesired behavior, we theoretically investigate evidential models and identify a fundamental limitation that explains the inferior performance: existing evidential activation functions create zero evidence regions, which prevent the model to learn from training samples falling into such regions. A deeper analysis of evidential activation functions based on our theoretical underpinning inspires the design of a novel regularizer that effectively alleviates this fundamental limitation. Extensive experiments over many challenging real-world datasets and settings confirm our theoretical findings and demonstrate the effectiveness of our proposed approach.
AAAI 2023

Evidential Conditional Neural Processes

Deep Shankar Pandey, and Qi Yu

In Proceedings of the Thirty-Seventh AAAI Conference on Artificial Intelligence and Thirty-Fifth Conference on Innovative Applications of Artificial Intelligence and Thirteenth Symposium on Educational Advances in Artificial Intelligence 2023

Abs

The Conditional Neural Process (CNP) family of models offer a promising direction to tackle few-shot problems by achieving better scalability and competitive predictive performance. However, the current CNP models only capture the overall uncertainty for the prediction made on a target data point. They lack a systematic fine-grained quantification on the distinct sources of uncertainty that are essential for model training and decision-making under the few-shot setting. We propose Evidential Conditional Neural Processes (ECNP), which replace the standard Gaussian distribution used by CNP with a much richer hierarchical Bayesian structure through evidential learning to achieve epistemic-aleatoric uncertainty decomposition. The evidential hierarchical structure also leads to a theoretically justified robustness over noisy training tasks. Theoretical analysis on the proposed ECNP establishes the relationship with CNP while offering deeper insights on the roles of the evidential parameters. Extensive experiments conducted on both synthetic and real-world data demonstrate the effectiveness of our proposed model in various few-shot settings.
CVPR 2022

Multidimensional Belief Quantification for Label-Efficient Meta-Learning

Deep Shankar Pandey, and Qi Yu

In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Jun 2022

Abs

Optimization-based meta-learning offers a promising direction for few-shot learning that is essential for many real-world computer vision applications. However, learning from few samples introduces uncertainty, and quantifying model confidence for few-shot predictions is essential for many critical domains. Furthermore, few-shot tasks used in meta training are usually sampled randomly from a task distribution for an iterative model update, leading to high labeling costs and computational overhead in meta-training. We propose a novel uncertainty-aware task selection model for label efficient meta-learning. The proposed model formulates a multidimensional belief measure, which can quantify the known uncertainty and lower bound the unknown uncertainty of any given task. Our theoretical result establishes an important relationship between the conflicting belief and the incorrect belief. The theoretical result allows us to estimate the total uncertainty of a task, which provides a principled criterion for task selection. A novel multi-query task formulation is further developed to improve both the computational and labeling efficiency of meta-learning. Experiments conducted over multiple real-world few-shot image classification tasks demonstrate the effectiveness of the proposed model.