iiinova — where your digital self keeps evolving

Harit Vishwakarma

@harit7.bsky.social

25 followers 20 following 8 posts

Ph.D. Candidate at UW-Madison harit7.github.io

Posts Replies Media

Reposted by Harit Vishwakarma

Agents4Academia @agents4academia.bsky.social · 10/07/2026

First up a hackathon: 20+ researchers · two weeks · @oxfordstatistics.bsky.social × @nuscomputing.bsky.social × @NTUSingapore and tokens on tap by @anthropic.com. Outcome: Five open‑source AI agents solving common workflow frustrations across the whole research lifecycle. 🧵

255

Reposted by Harit Vishwakarma

Agents4Academia @agents4academia.bsky.social · 10/07/2026

Hello world 👋 Agents4Academia is a community‑led effort to explore and build open‑source agents for academic and research work — by researchers, for researchers. agents4academia.org

061

Harit Vishwakarma @harit7.bsky.social · 11/12/2024

@srinathnamburi.bsky.social

010

Harit Vishwakarma @harit7.bsky.social · 11/12/2024

Join us in the evening poster session (#1906) to learn more about it and chat about auto-labeling and data-centric AI. Thanks to the amazing co-authors: Yi (Reid) Chen, Sui Jiet Tay, Srinath Namburi, @fredsala.bsky.social, Ramya Korlakai Vinayak.

120

Harit Vishwakarma @harit7.bsky.social · 11/12/2024

Our method learns confidence functions tailored for efficient and reliable auto-labeling. Using these in TBAL boosts the no. of auto-labeled points by up to 60% (while making < 5% auto-labeling errors) compared to baselines like softmax and several training-time and post-hoc calibration techniques.

110

Harit Vishwakarma @harit7.bsky.social · 11/12/2024

Introducing Colander, our framework for learning optimal confidence functions for TBAL! We formulate the auto-labeling objective as an optimization problem over the space of confidence functions and thresholds.

100

Harit Vishwakarma @harit7.bsky.social · 11/12/2024

We systematically study the limitations of popular confidence functions like softmax outputs and off-the-shelf calibration techniques. The result? Too few auto-labeled points or large auto-labeling errors.

100

Harit Vishwakarma @harit7.bsky.social · 11/12/2024

The choice confidence function is crucial in TBAL – if it's not aligned with the auto-labeling objective, it can be detrimental to performance. We show commonly used confidence functions fall short.

100

Harit Vishwakarma @harit7.bsky.social · 11/12/2024

TBAL is a promising auto-labeling technique. It iteratively acquires human labels for small data chunks, trains a model, and auto-labels points where the model's confidence is above a threshold. The goal? Maximize coverage (proportion of auto-labeled points) with bounded auto-labeling error.

100

Harit Vishwakarma @harit7.bsky.social · 11/12/2024

Excited to present Colander at #NeurIPS2024, our new framework for optimizing confidence functions to make auto-labeling more efficient and reliable. Check out our poster #1906 at today's evening poster session. Wed, Dec 11, 4:30–7:30 p Poster #1906 Project: harit7.github.io/colander

142