Fwd: [Wiki-research-l] [Wikimedia Research Showcase] May 20, 2020: Human in the Loop Machine Learning

classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view

Fwd: [Wiki-research-l] [Wikimedia Research Showcase] May 20, 2020: Human in the Loop Machine Learning

Pine W

( https://meta.wikimedia.org/wiki/User:Pine )

---------- Forwarded message ---------
From: Janna Layton <[hidden email]>
Date: Fri, May 15, 2020 at 8:05 PM
Subject: [Wiki-research-l] [Wikimedia Research Showcase] May 20, 2020:
Human in the Loop Machine Learning
To: <[hidden email]>,
<[hidden email]>,
<[hidden email]>

Hi all,

The next Research Showcase will be live-streamed on Wednesday, May 20, at
9:30 AM PDT/16:30 UTC.

This month we will learn about recent research on machine learning systems
that rely on human supervision for their learning and optimization -- a
research area commonly referred to as Human-in-the-Loop ML. In the first
talk, Jie Yang will present a computational framework that relies on
crowdsourcing to identify influencers in Social Networks (Twitter) by
selectively obtaining labeled data. In the second talk, Estelle Smith will
discuss the role of the community in maintaining ORES, the machine learning
system that predicts the quality in Wikipedia applications.

YouTube stream: https://www.youtube.com/watch?v=8nDiu2ebdOI

As usual, you can join the conversation on IRC at #wikimedia-research. You
can also watch our past research showcases here:

This month's presentations:

*OpenCrowd: A Human-AI Collaborative Approach for Finding Social
Influencers via Open-Ended Answers Aggregation*

By: Jie Yang, Amazon (current), Delft University of Technology (starting

Finding social influencers is a fundamental task in many online
applications ranging from brand marketing to opinion mining. Existing
methods heavily rely on the availability of expert labels, whose collection
is usually a laborious process even for domain experts. Using open-ended
questions, crowdsourcing provides a cost-effective way to find a large
number of social influencers in a short time. Individual crowd workers,
however, only possess fragmented knowledge that is often of low quality. To
tackle those issues, we present OpenCrowd, a unified Bayesian framework
that seamlessly incorporates machine learning and crowdsourcing for
effectively finding social influencers. To infer a set of influencers,
OpenCrowd bootstraps the learning process using a small number of expert
labels and then jointly learns a feature-based answer quality model and the
reliability of the workers. Model parameters and worker reliability are
updated iteratively, allowing their learning processes to benefit from each
other until an agreement on the quality of the answers is reached. We
derive a principled optimization algorithm based on variational inference
with efficient updating rules for learning OpenCrowd parameters.
Experimental results on finding social influencers in different domains
show that our approach substantially improves the state of the art by 11.5%
AUC. Moreover, we empirically show that our approach is particularly useful
in finding micro-influencers, who are very directly engaged with smaller

Paper: https://dl.acm.org/doi/fullHtml/10.1145/3366423.3380254

*Keeping Community in the Machine-Learning Loop*

By:  C. Estelle Smith, MS, PhD Candidate, GroupLens Research Lab at the
University of Minnesota

On Wikipedia, sophisticated algorithmic tools are used to assess the
quality of edits and take corrective actions. However, algorithms can fail
to solve the problems they were designed for if they conflict with the
values of communities who use them. In this study, we take a
Value-Sensitive Algorithm Design approach to understanding a
community-created and -maintained machine learning-based algorithm called
the Objective Revision Evaluation System (ORES)—a quality prediction system
used in numerous Wikipedia applications and contexts. Five major values
converged across stakeholder groups that ORES (and its dependent
applications) should: (1) reduce the effort of community maintenance, (2)
maintain human judgement as the final authority, (3) support differing
peoples’ differing workflows, (4) encourage positive engagement with
diverse editor groups, and (5) establish trustworthiness of people and
algorithms within the community. We reveal tensions between these values
and discuss implications for future research to improve algorithms like


Janna Layton (she, her)
Administrative Assistant - Product & Technology
Wikimedia Foundation <https://wikimediafoundation.org/>
Wiki-research-l mailing list
[hidden email]

Wikipedia-l mailing list
[hidden email]