Fwd: Fully Funded PhD studentship: Robotic Language Acquisition & Dialogue Systems
5 views
Skip to first unread message
Camilo Thorne
Jan 6, 2021, 10:31:20 AM1/6/21
to sp...@googlegroups.com, ai...@aixia.it, logi...@dimap.ufrn.br
Camilo Thorne
"Exegi monumentum aere perennius"
(Horatius, Ode III-30)
---------- Forwarded message ---------
From: Frank Foerster <frank.f...@gmx.net>
Date: Wed, Jan 6, 2021 at 4:29 PM
Subject: Fully Funded PhD studentship: Robotic Language Acquisition & Dialogue Systems
To: <SIGAI-A...@listserv.acm.org>
From: Frank Foerster <frank.f...@gmx.net>
Date: Wed, Jan 6, 2021 at 4:29 PM
Subject: Fully Funded PhD studentship: Robotic Language Acquisition & Dialogue Systems
To: <SIGAI-A...@listserv.acm.org>
[Apologies for multiple copies due to cross-posting]
============
PhD studentship
ROBOTIC LANGUAGE ACQUISITION & DIALOGUE SYSTEMS
Adaptive Systems Research Group
Centre for Computer Science and Informatics Research
School of Engineering and Computer Science
University of Hertfordshire, UK
Contact: Frank Foerster (f.foe...@herts.ac.uk)
============
Short-listings will start on 29 January 2021*
Bursary £15,285 p.a.
* application before this date is strongly encouraged
We invite applications for a PhD studentship at the University of
Hertfordshire, UK, under the supervision of Dr. Frank Foerster in
the area of robotic language acquisition and potential
applications to dialogue systems. We are particularly interested in
pursuing research that connects to the following two topics, but
alternative ideas will be considered too.
Research Topics
Topic 1: Socially driven Machine Learning in Robotic Language
Acquisition
We have developed a language acquisition approach that strongly
relies on social cues similar to those present in child-directed speech
(CDS). The big advantage of this approach is that it can cope with
unconstrained robot-directed speech rather than relying on heavily
structured and constrained natural language input by eliciting vocal
behaviour from human participants similar to that found in CDS.
We have reasons to believe that the application of reinforcement
learning in its widest sense may lead to large improvements of the
core learning algorithm. This research topic involves areas ranging
from human-robot interaction, over machine learning,
and social signal processing, to affective computing.
Topic 2: Linking Robotic Language-Acquisition Research with
Dialogue Systems
Historically robotic language acquisition research has been mostly
restricted to developmental robotics and similarly small research
areas. More recently, the larger computer vision and dialogue
systems communities have picked up on some of the ideas such as
symbol grounding. So far, these approaches focused largely on
the grounding ( = attaching meaning to) of referential words such as
object labels, tangible actions and events. Part of our past research
has focused on the grounding of non-referential words, including
negation words, relating to “non-tangible” concepts such as affect
or volition. We believe that dialogue systems could benefit from
approaches in developmental robotics that focus on such
“non-referential” utterances, and invite the prospective student to
explore this space.
Independent of the particular topic, we have recently become member
of the HomeBank corpora, and would strongly encourage interaction and
collaboration with developmental psychologists and psycholinguists.
Person Profile
You will have an excellent first degree and a very keen interest and
motivation in language or language acquisition. Optimally you
should have an excellent background in Computer Science,
Computational/Cognitive Robotics, (computational) linguistics,
artificial intelligence, or similar disciplines with a considerable
quantitative/computational component.
Due to the interdisciplinary nature of the topic we will also consider
applicants with a background in (developmental) psychology,
philosophy, or pragmatics as long as you have some experience
in programming, machine learning, or dialogue systems.
Prior experience with topics such as reinforcement learning, or
statistical learning more generally is highly desirable, but not
essential if the quantitative background is otherwise very strong.
The knowledge of later Wittgenstein is a big plus.
More information including relevant papers are provided in the
following PDF:
If you have questions or have alternative interested in applying,
please contact
Dr. Frank Foerster
============
Research in Computer Science at the University of Hertfordshire has
been recognised as excellent by the latest Research Assessment
Exercise, with 50% of the research submitted being rated as
internationally excellent or world leading.
The University of Hertfordshire provides a very stimulating
environment, offering a large number of specialised and
interdisciplinary seminars as well as general training and researcher
development opportunities. The University is situated in Hatfield, in
the green belt just north of London. Hatfield is close to Central
London (less than 25 minutes by direct train to Kings Cross), with
convenient access to Stansted, Luton and Heathrow airports, and, via
the nearby historic town of St. Albans, also to Gatwick airport.
Application forms and submission instructions are available under
Please note: You must download the application form to your computer
before you complete it. If you complete the form in the browser
window, the information you have entered may be lost when the form is
saved.
Eligibility
-----------
Applicants must submit or obtain:
- a completed application form, which can be obtained from website linked above
- a research statement stating past experience/research background and
interest and motivation/research ambition for current project
- two academic references
- copies of qualification certificates and transcripts
- a copy of your passport
Applications from outside the UK or EU are eligible for this studentship.
The application form should be returned to:
Have an AI-relevant announcement? Send it to sigai-a...@listserv.acm.org
To unsubscribe from the SIGAI-ANNOUNCE list, click the following link:
https://LISTSERV.ACM.ORG/SCRIPTS/WA-ACMLPX.CGI?SUBED1=SIGAI-ANNOUNCE&A=1
Reply all
Reply to author
Forward
0 new messages