Contact Us

Give us a call or drop by anytime, we endeavor to answer all inquiries within 24 hours.

map

Find us

PO Box 16122 Collins Street West Victoria, Australia

Email us

info@domain.com / example@domain.com

Phone support

Phone: + (066) 0760 0260 / + (057) 0760 0560

Lily Weng and Berk Ustun Receive $1.2M Grant to Advance Safety in Machine Learning

  • By HDSIComm
  • February 14, 2024
  • 497 Views

Lily Weng and Berk Ustun, faculty with the Halıcıoğlu Data Science Institute at the University of California San Diego (UCSD), have been awarded a $1.2 million research grant from the National Science Foundation (NSF). The grant, part of the NSF’s Medium program, is for their research project “Foundations of Recourse Verification in Machine Learning.”

In today’s landscape, machine learning algorithms wield immense power, shaping pivotal decisions across industries like lending, hiring, and public service allocation. However, a critical flaw persists: these models often overlook “actionability,” disregarding individuals’ ability to influence the features informing their predictions. Such oversight can have dire repercussions, trapping individuals in cycles of denied credit or employment opportunities due to fixed model predictions.

Weng and Ustun’s research endeavors to rectify this issue through the innovative concept of “recourse verification.” This entails devising a formal verification protocol ensuring machine learning models alter their predictions through actionable adjustments in feature space. he PIs aim to craft adaptable tools capable of evaluating and refining models to guarantee recourse for every individual by embedding these constraints into optimization problems.f successful, this research could revolutionize the field of machine learning by providing robust solutions to ensure access in consumer-facing applications.

This research addresses a significant vulnerability in machine learning models used in critical domains like lending and hiring – the potential for models to deny access by generating predictions that individuals cannot change. Given the regulatory and ethical considerations surrounding such applications, the outcomes of this research will inform policy and regulation design aimed at guaranteeing access and consumer protection.

This research holds immense promise in not only advancing the field of machine learning but also promoting safety, fairness and accessibility in decision-making processes that profoundly shape individuals’ lives.