Mechanisms and processes for operationalizing responsible ML

As Machine Learning (ML) is being embedded into more services, the need to address broader business and social concerns about its use is growing. It is not enough to have an accurate model or at least having accuracy as its sole quality criterion. It is necessary for models to implement multi-dimensional quality criteria and be defensible with all stakeholders and build trust. ML models have to be both economically viable and FAccT (Fair, Accountable, Transparent). The terminology is new but not the need to defend models or to attest they can be trusted. Such requirements were present from the 70s for credit scoring models. What has changed is the scale and scope. In this session, we will explore:

  1. Current status of the FAccT conversation and methods (legal, consumer expectations etc.)
  2. What are the compelling reasons for an organisation to invest in FAccT? What it takes to achieve end-to-end FAccT?
  3. Practical challenges in implementing FAccT both at organisational and at technical level
  4. Various efforts and initiatives around the world (academic, industry-related, products/tooling)
  5. What to anticipate in the next year or two

Panelists
Dr. Joshua Rubin, Data Scientist, Fiddler Labs, San Francisco
Dr. Yiannis Kanellopoulos, Founder, Code4Thought, Athens

Notes

  1. Fiddler Webinar: Responsible AI Panel Discussion: Hype vs Reality & Building Responsibly at Every Stage of AI
  2. The Case for Process Fairness in Learning
  3. The Human Body is a Black Box: Supporting Clinical Decision-Making with Deep Learning

Registration Register to participate via Zoom. Zoom link will be shared one day before the event. Or, watch the livestream on this page.

Registered participants can also leave comments and questions for the hosts and speakers, which will be taken up during the session.

Previous session: The previous session was held on 15 July. Summary of the session is available on https://hasgeek.com/fifthelephant/making-data-science-work-5/

and Indrayudh Ghoshal of Scribble Data will host the meetup. Scribble Data is a Bangalore/Toronto startup, active in the data community. Scribble is a MLOps product company whose main offering is ML Feature Store that increases trust and efficiency of production models.

About the curators: Suchana Seth, Venkata Pingali and Indrayudh Ghoshal of Scribble Data have curated this session. Suchana Seth is Fellow at Berkman Klein Center, Harvard University. Scribble Data is a Bangalore/Toronto startup, active in the data community.

About the series producer: The Fifth Elephant is platform for practitioners working with data (engineering, to application of data science for different use cases) to showcase their work and to collaborate.

For further inquiries, contact 7676332020 or write to fifthelephant.editorial@hasgeek.com

Videos

See all
Operationalizing Responsible ML

Operationalizing Responsible ML

1 hour29 July 2020

Hosted by

Jump starting better data engineering and AI futures