Program | Speakers | Registration | Organizers | Venue

The first Human-aligned AI Summer School (HAAISS) will be held in Prague from 2nd to 5th August. The focus of the first year will be on "learning from humans," including subtopics such as inverse reinforcement learning, modeling bounded rational agents and latest trends in AI alignment research.

Format of the school

The school is focused on teaching approaches and frameworks, less on presentation of latest research results. The content of the school is mostly technical - it is assumed the attendees understand current ML approaches such as deep learning. The intended audience of the school are researchers interested in learning more about the AI alignment topics, PhDs, researchers working in ML/AI outside academia, and talented students.

Program (beta)

Thu 2nd
Morning session
Venue: Faculty of Mathematics and Physics

9:00-10:00 Registration
10:00-11:00 Opening session

11:00-11:20 Coffee break
11:20-12:00 Overview of the school program (Jan Kulveit)

Lunch & afternoon session.
Venue: Snemovni 7 event space

12:30-14:00 Lunch (catered)

14:00-15:30 Value learning (Daniel Filan)
15:30-15:50 Coffee break
16:00-17:00 Value learning (Daniel Filan)
17:00-18:00 Probability, utility and Jeffrey-Bolker rotation (Abram Demski)

Evening program:
Venue: Snemovni 7

19:00-21:30 Welcome reception

Fri 3rd
Venue: Snemovni 7 event space

9:30-11:00 Bounded rationality (Daniel Braun)
11:00-11:20 Coffee break
11:20-12:30 Bounded rationality (Daniel Braun)

12:30-13:30 Lunch (catered)

13:30-14:30 Bounded rationality (Filip Matějka)
14:30-15:00 Coffee break
15:30-17:00 Bounded rationality (Daniel Filan)
17:00-17:20 Coffee break
17:20-18:00 Fireside chat

(Dinner individually)

Evening program
Baroque refectory of the Dominican cloister

20:00-22:00 Cello concerto (light refreshments will be served after the concert)

Sat 4th
Venue: Snemovni 7 event space

9:30-10:0 Lightning talks (early career researchers)

11:00-11:20 Coffee break
11:20-12:30 Impact measures and side effects (Victoria Krakovna)

12:30-14:00 Lunch (catered)

14:00-15:40 Embedded agents (Abram Demski)
15:40-16:00 Coffee break
16:00-17:00
17:00-18:00 Lightning talks (early career researchers) Evening program
Venue: Lucerna

19:00-22:00 School dinner on the roof of Lucerna palace

Sun 5th
Venue: Snemovni 7 event space

9:30-11:00 Strategic considerations in AI development (Miles Brundage)
11:00-11:20 Coffee break
11:20-12:40 Panel discussion (Miles Brundage,Victoria Krakovna,Daniel Filan,Abram Demski)

12:40-14:20 Lunch (catered)

14:20-14:50 Closing session (Jan Kulveit)

Speakers

Daniel Alexander Braun
Professor of Learning Systems, Ulm University

Daniel Braun is Professor of Learning Systems at Ulm University. He habilitated in neural and behavioral biology and cognitive science at the Eberhard Karls Universität Tübingen. Previously, he was a visiting PhD student and postdoctoral research associate in the Computational and Biological Learning Laboratory Cambridge University and a visiting scientist in the Computational Learning and Motor Control Laboratory at the University of Southern California. In 2011 he was awarded an Emmy-Noether-fellowship by the Deutsche Forschungsgemeinschaft to establish the independent research group "Sensorimotor Learning and Decision-making" at the Max-Planck-Institutes for Biological Cybernetics and Intelligent Systems in Tübingen. In 2015 he was awarded an ERC Starting Grant "BRISC: Bounded rationality in sensorimotor coordination". He has doctorates in natural science and philosophy from the Albert-Ludwigs-Universität, Freiburg, in the subject areas of computational neuroscience and philosophy of mind respectively.

His research interests include cognitive modelling, decision-making and bounded rationality,a abstraction and structural learning, sensorimotor learning and control, learning robots and biological information processing


Filip Matějka
Associate Professor with Tenure, CERGE-EI
Research Fellow, Center for Economic and Policy Research, London

Filip Matějka is an Associate Professor with Tenure at CERGE, Charles University, Prague, and a Researcher at the Economics Institute of the Czech Academy of Sciences. He obtained his Ph.D in applied mathematics from Princeton University and M.A. in Physics from Charles University in Prague.. Previously, he was a visiting Researcher, UC Berkeley, and visiting scholar at New York University. He was awarded ERC Starting Grant "Behavioral and Policy Implications of Rational Inattention" and in 2016 won The Neuron Award for promising young scientists.

His research areas are information economics, theory of rational inattention, behavioral economics and macroeconomics.


Miles Brundage
Research Fellow, Future of Humanity Institute, Oxford University

Miles Brundage is a Research Fellow at the Future of Humanity Institute, Oxford University. He has previously worked at the International Security and Biopolicy Center, University of Oklahoma, Belfer Center for Science and International Affairs, Harvard University, and Center for Energy and Environmental Research in the Human Sciences, Rice University.

His research focuses on the societal and strategic implications of advances in artificial intelligence, including interpretability, capability assessment, and expert elicitation.


Victoria Krakovna
Research Scientist, DeepMind

Victoria Krakovna is a Research Scientist at DeepMind, working on AI safety. Her research focuses on avoiding negative side effects of AI systems and preventing reward hacking. She has also created the AI Impacts project, a resource for information about the likely societal impacts of AI. She has a PhD in Statistics from Harvard University and a BA in Mathematics and Computer Science from UC Berkeley.


Daniel Filan
DPhil Student, Oxford University

Daniel Filan is a DPhil student at Oxford University working on AI safety. He has previously worked as a research assistant for Stuart Russell at UC Berkeley.


Abram Demski
Research Fellow, MIRI

Abram Demski is a Research Fellow at Machine Intelligence Research Institute (MIRI), working on AI safety. His current primary research areas are decision theory and embedded agency. He received a PhD in Computer Science from the University of Southern California.

Registration and fees

Regular school fee is € 200. Student fee is € 100. Thanks to our sponsors, limited financial assistance, including partial travel costs reimbursement, is available for participants who want to work on AI alignment research but travel or registration costs would prevent them from attending the school.

Organizers

The conference is organized by

Program: Jan Kulveit (main coordinator), Tomáš Gavenčiak,Jan Romportl
Operations: Hana Kalivodova

Venues

Faculty of Mathematics and Physics, Charles University
Malostranske náměstí 25, Praha 1
First floor - the way from the building entrance will be signposted.

Sněmovní 7 event & coworking space
Sněmovní 7, Praha 1

Baroque refectory of the Dominican cloister
Jilská 7, Praha 1

Rooftop of the Lucerna Palace
Štěpánská 61, Praha 1

Sponsors

Kiwi.com