Karan Bania

News

• 09/24 - Got a placement offer from Gan.AI!
• 08/24 - Made the President of SAiDL for the year 2024-25.
• 08/24 - Our work on MolGAN and Normalizing flows is out on arXiv [Link].
• 07/24 - I am a DataThon volunteer at IndoML'24, The contest goes live on 5th August, register here!
• 07/24 - A SAiDL blogpost led by me, on Graphormer was accepted to the GRaM Workshop at ICML'24. It is live here.
• 06/24 - Our reproduction of Teaching CLIP to Count to Ten is on arXiv [Link].
• 05/24 - Started my DAAD WISE Internship.
• 04/24 - Won first prize in ACM Goa Chapter's event PitchIt! organized at BITS Goa for my research presentation. You can check out the slides here.
• 02/24 - Our Spring Induction assignment for the 2024-25 cycle is out! [Link] Fill out this form and join the Slack workspace for any doubts. Looking forward to your submissions!
• 02/24 - Selected for the DAAD WISE program (one of 250 undergraduates from India) at RUB, Germany for summer 2024! Institut Für Neuroinformatik will be hosting me.
• 01/24 - Selected for the MITACS GRI program (over 30,000 applicants) at the Ontario Tech University, Canada for summer 2024! VC Lab will be hosting me. (I rejected the offer)
• 01/24 - TAing the BITS F464 - Machine Learning course at BITS Goa this semester. (You can check out our labs here)
• 01/24 - Made the General Secretary of SAiDL for the year 2023-24.
• 12/23 - Started contributing to DeepChem.
• 10/23 - Check out our annual AI Symposium! Register here. [linktree]
• 08/23 - Started working at APPCAIR as a student researcher.
• 08/23 - TAing the CS F214 - Logic in Computer Science course at BITS Goa this semester.
• 07/23 - One of four people from our batch to get inducted into SAiDL!
• 06/23 - Selected for IIITH's 7th Summer School on AI with focus on Computer Vision.
• 05/23 - Selected for Neuromatch Summer School in the Deep Learning Track.

Projects & Publications

• APPCAIR I

  This project's report is under review at a double-blind venue. Details soon.

• APPCAIR II / Radio-LM

  Working on a novel pseudo replacement to RLHF through prompting. This project also involves creating a modified model with this pseudo-RLHF and then uing it to find a Human-window of LLM explanation-understandability, in corroboration with Donald Michie's work. We are testing this theory on Radiology/Med students. This work's presentation led to the first prize in ACM event, slides.
  Results on the way ;)

• DeepChem

  Contributing to the DeepChem library. As of now I have contributed to the Introduction to Materials Science Tutorial. Also co-developed MolGAN and Normalizing Flow implementations in PyTorch. The abstract from that work from here has been accepted at MoML 2024. I also developed the tutorials for them. [MolGAN] [NormFlows] [Preprint] I am currently working on an "Equivariant Modelling in Materials Science" tutorial. This project's report is under review at another double-blind venue. More details soon.

• [RE] Teaching CLIP to Count to Ten

  This is a student-led reproduction of the Google Brain paper Teaching CLIP to Count to Ten. We developed novel (incremental) augmentations to the loss function, which showed improvement in our small scale experiments. Apart from this, we created data processing pipelines and also made their evaluation dataset completely public. This work has been submitted to ReScience C, [Preprint]

• Event Stream analysis using Deep SSMs

  This project's report is under review at a double-blind venue. Details soon.

Acknowledgement