Skip to main content

Adaptive conversational agents – (partly) using a crowd-in-the-loop paradigm

Time: Fri 2020-11-06 15.15

Location: This seminar will be fully virtual

Participating: Patrik Jonell

Export to calendar

In this presentation I will be talking about what I have been been doing
during the last three and a half years as a PhD student. My research is
in interlocutor-aware conversational agents, i.e. agents which adapt
their behavior based on the behavior of the interlocutor using
data-driven methods. I do this, in part, using a so-called
crowd-in-the-loop paradigm, where I involve crowdworkers in the
processes of developing these systems.

This work has identified and explored methods in three connected key
areas which are crucial to developing such adaptive agents; data
collection, machine learning, and evaluation. The crowdworkers are
involved in both the data collection process and used to evaluate the
outcome of the machine learning methods. Based on the outcome of the
evaluations, new data might be collected or the machine learning methods
changed. This process is then performed iteratively until the desired
outcome is achieved.

In this talk I will delve deeper into each of these three areas,
primarily focusing on the work that has been done since the last
seminar. I will also give a brief overview of previous work to provide
some context, and discuss some future directions I'm planning to explore.

--------------

Zoom URL

Other ways to join:
Meeting ID: 695 7354 2160
One tap mobile
+46844682488,,69573542160# Sweden
+46850163827,,69573542160# Sweden
Dial by your location
  +46 8 4468 2488 Sweden
  +46 8 5016 3827 Sweden
  +46 8 5050 0828 Sweden
  +46 8 5050 0829 Sweden
  +46 8 5052 0017 Sweden
  +46 850 539 728 Sweden
Meeting ID: 695 7354 2160
Find your local number: https://kth-se.zoom.us/zoomconference
Join by SIP
69573542160@zoom.nordu.net
 From a KTH Cisco Video System you can just input 69573542160 and then
the CALL-button
Join by H.323
109.105.112.236
109.105.112.235
Meeting ID: 695 7354 2160