Skip to content
View shubhamOjha1000's full-sized avatar
4 followers · 31 following

Block or report shubhamOjha1000

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Please don't include any personal information such as legal names or email addresses. Maximum 100 characters, markdown supported. This note will be visible to only you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
shubhamOjha1000/README.md

Hi there!

IIT KGP logo   Omdena logo   IIT Patna logo   coe ai logo

I am Shubham Ojha, currently research intern at IIT Kharagpur(KLIV research group). In the past, I had the fortune of interning at Tech Mahindra AI Lab,Omdena & IIT Patna working on wide spectrum research verticals of AI.

  • My interests are quite wide as I strongly belive that rather than going deep in one domain its imp to increase the breadth:

    • How to go beyon IID hypothesis i.e. to develope algorithms which are roboust to OOD. I belive that in order to go beyond IID hypothesis, we need to do following things :- 1. Induce some amount of prior in the model, 2. Need to introduce concept of agent-environment so that models learns by interacting with env a robust represenation(disentangle the various underlying explanatory factors of variations).
    • AI for natural sciences. Interested in the field of bioinformatics(such as genomics, 3D structure modelling of Proteins, biomedical imaging, novel drug desing) and climate change(how to use data captured from the env and fuse it with domain laws to make predictions).
    • Aligning AI with common consesnsus of human values
    • AI for Education
    • How to quantify uncertanity in DL. Current DL models are highly confiednt when they are wrong, this behavior is undesirable. DL models should be able to correctly quantify the amount of unceratnity in thier predictions. I am interested in probablistic approach of ML to get the desired uncertanity.
    • Building underlying core AI technologies upon which many AI powered products and services can be built. For eg Query adaptive video indexing can be thought of as an AI powered service and scene understanding can be thought of as an underlying core AI technology.

  • For any projects,ideas,discussions or collaborations feel free to reach me at my mail ✉️ : shubham.ojha1000@gmail.com

Pinned Loading

  1. Conditional_Covg_CP_code_AAAI_SA Conditional_Covg_CP_code_AAAI_SA Public

    Python

  2. Optimizing-Conformal-Prediction-Sets-for-Pathological-Image_Classification Optimizing-Conformal-Prediction-Sets-for-Pathological-Image_Classification Public

    Python

  3. mitigating_feature_bias_WIML_NIPS_2024 mitigating_feature_bias_WIML_NIPS_2024 Public

    Python

  4. UQ-in-DL-Models-for-Cervical-Cytology UQ-in-DL-Models-for-Cervical-Cytology Public

    Uncertainty Quantification in DL Models for Cervical Cytology

    Python 1