University of Cambridge > Talks.cam > Engineering Safe AI

  Browse A-Z

Log in

University Account External (via Google)

πŸ” Log In

Information on

Subscribing to talksdetails Finding a talkdetails Adding a talkdetails Disseminating talksdetails Help and Documentationdetails

ℹ️ Information

Engineering Safe AI

Further detail
Subscribe using ical/vcal (Help)

Presentations and discussions about possible solutions to the value alignment problem.

If you have a question about this list, please contact: AdriΓ  Garriga Alonso. If you have a question about a specific talk, click on that talk to find its organiser.

0 upcoming talks and 38 talks in the archive.

Engineering Safe AI: Robert Miles

UserRobert Miles.

HousePavillion Room, Hughes Hall, University of Cambridge, Cambridge, CB1 2EW.

ClockWednesday 29 January 2020, 19:00-20:30

Can Machines Read our Minds?

Starting time 30min later than usual

UserBrier Rigby Dames (University of Cambridge).

HouseEngineering Department, CBL Seminar room BE4-38.

ClockWednesday 12 June 2019, 17:30-19:00

How useful is quantilization for mitigating specification-gaming?

UserSpeaker to be confirmed.

HouseEngineering Department, CBL Seminar room BE4-38.

ClockWednesday 22 May 2019, 17:00-18:30

Misleading meta-objectives and hidden incentives for distributional shift

UserPaolo Bova (University of Cambridge).

HouseEngineering Department, CBL Seminar room BE4-38.

ClockWednesday 08 May 2019, 17:00-19:00

Causal Reasoning from Meta-reinforcement Learning

UserJakub Perlin (University of Cambridge).

HouseEngineering Department, CBL Seminar room BE4-38.

ClockWednesday 13 March 2019, 17:00-19:00

Inverse Game Theory

UserGabija MarΕ‘alka.

HouseEngineering Department, CBL Seminar room BE4-38.

ClockWednesday 06 March 2019, 17:00-19:00

Goals vs Utility Functions

UserAdriΓ  Garriga Alonso (University of Cambridge).

HouseEngineering Department, CBL Seminar room BE4-38.

ClockWednesday 27 February 2019, 17:00-19:00

Who do we want to control human-level AI?

UserJade Leung (Center for the Governance of AI, University of Oxford).

HouseBoys Smith Room, Fisher Building, St John's College.

ClockFriday 22 February 2019, 19:30-21:00

Bayesian Theory of Mind: Modeling Joint Belief-Desire Attribution

UserEdward Young (University of Cambridge).

HouseEngineering Department, CBL Seminar room BE4-38.

ClockWednesday 20 February 2019, 17:00-19:00

Ambitious Value Learning

UserAdriΓ  Garriga Alonso (University of Cambridge).

HouseCambridge University Engineering Department, CBL Seminar room BE4-38.

ClockWednesday 13 February 2019, 17:00-19:00

Machine Theory of Mind

UserPaolo Bova (University of Cambridge).

HouseCambridge University Engineering Department, CBL Seminar room BE4-38.

ClockWednesday 06 February 2019, 17:00-19:00

Embedded Agency

UserAdriΓ  Garriga Alonso (University of Cambridge).

HouseCambridge University Engineering Department, CBL Seminar room BE4-38.

ClockWednesday 30 January 2019, 17:00-19:00

Comprehensive AI Services

UserAdriΓ  Garriga Alonso (University of Cambridge).

HouseCambridge University Engineering Department, CBL Seminar room BE4-38.

ClockWednesday 23 January 2019, 17:00-19:00

Incomplete Contracting and AI Alignment

UserPaolo Bova (University of Cambridge).

HouseCambridge University Engineering Department, CBL Seminar room BE4-38.

ClockWednesday 28 November 2018, 17:00-19:00

The Algorithmic Foundations of Differential Privacy (Chapters 1 and 2)

UserJames Bell, University of Cambridge.

HouseCambridge University Engineering Department, CBL Seminar room BE4-38.

ClockWednesday 21 November 2018, 17:00-19:00

Dynamic Safe Interruptibility for Decentralized Multi-Agent Reinforcement Learning

UserAdriΓ  Garriga Alonso (University of Cambridge).

HouseCambridge University Engineering Department, CBL Seminar room BE4-38.

ClockWednesday 14 November 2018, 17:00-19:00

Measuring and avoiding side effects using relative reachability

UserAdriΓ  Garriga Alonso (University of Cambridge).

HouseCambridge University Engineering Department, CBL Seminar room BE4-38.

ClockWednesday 07 November 2018, 17:00-19:00

Interpretable Machine Learning

UserTameem Adel (University of Cambridge).

House Cambridge University Engineering Department, CBL Seminar room BE4-38. See https://www.openstreetmap.org/#map=18/52.19804/0.11969.

ClockWednesday 31 October 2018, 17:00-18:30

Scaling inverse reinforcement learning for human-compatible AI

UserAdam Gleave, UC Berkeley.

House Cambridge University Engineering Department, CBL Seminar room BE4-38. See https://www.openstreetmap.org/#map=18/52.19804/0.11969.

ClockTuesday 23 October 2018, 17:00-18:30

Motivation for this group, Goodhart's Law

UserJames Bell, University of Cambridge.

House Cambridge University Engineering Department, CBL Seminar room BE4-38. See https://www.openstreetmap.org/#map=18/52.19804/0.11969.

ClockWednesday 17 October 2018, 17:00-18:30

Approaches to avoiding negative side effects

UserAdriΓ  Garriga Alonso (University of Cambridge).

House Cambridge University Engineering Department, CBL Seminar room BE4-38. For directions see http://learning.eng.cam.ac.uk/Public/Directions.

ClockWednesday 30 May 2018, 17:00-18:30

AI Safety via Debate

UserBeth Barnes (University of Cambridge).

House Cambridge University Engineering Department, CBL Seminar room BE4-38. For directions see http://learning.eng.cam.ac.uk/Public/Directions.

ClockWednesday 09 May 2018, 17:00-18:30

AI risk strategy

UserRajashree Agrawal (University of Oxford).

House Cambridge University Engineering Department, CBL Seminar room BE4-38. For directions see http://learning.eng.cam.ac.uk/Public/Directions.

ClockWednesday 14 March 2018, 17:00-18:30

AI Safety Gridworlds: Is my agent 'safe'?

UserJessica Yung (University of Cambridge).

House Cambridge University Engineering Department, CBL Seminar room BE4-38. For directions see http://learning.eng.cam.ac.uk/Public/Directions.

ClockWednesday 28 February 2018, 17:00-18:30

Logical Induction: a computable approach to logical non-omniscience

UserAdriΓ  Garriga Alonso (University of Cambridge).

House Cambridge University Engineering Department, CBL Seminar room BE4-38. For directions see http://learning.eng.cam.ac.uk/Public/Directions.

ClockWednesday 21 February 2018, 17:00-18:30

Decision Boundary Geometries and Robustness of Neural Networks

UserSven Wang (University of Cambridge).

House Cambridge University Engineering Department, CBL Seminar room BE4-38. For directions see http://learning.eng.cam.ac.uk/Public/Directions.

ClockWednesday 14 February 2018, 17:00-18:30

Decision Theory for AI safety

UserRichard Ngo (University of Cambridge).

House Cambridge University Engineering Department, CBL Seminar room BE4-38. For directions see http://learning.eng.cam.ac.uk/Public/Directions.

ClockWednesday 07 February 2018, 17:00-18:30

Safe Exploration in Reinforcement Learning

UserFrances Ding (University of Cambridge).

House Cambridge University Engineering Department, CBL Seminar room BE4-38. For directions see http://learning.eng.cam.ac.uk/Public/Directions.

ClockWednesday 31 January 2018, 17:00-18:30

Amplification and dialogue as mechanisms for safe advanced AI

UserBeth Barnes, Computer Lab, University of Cambridge.

House Cambridge University Engineering Department, CBL Seminar room BE4-38. For directions see http://learning.eng.cam.ac.uk/Public/Directions.

ClockWednesday 24 January 2018, 17:00-18:30

Last term summary + discussion of topic importance

UserAdriΓ  Garriga Alonso (University of Cambridge).

House Cambridge University Engineering Department, CBL Seminar room BE4-38. For directions see http://learning.eng.cam.ac.uk/Public/Directions.

ClockWednesday 17 January 2018, 17:00-18:30

Counterargument to CIRL, and Safely Interruptible Agents

UserAdriΓ  Garriga Alonso (University of Cambridge).

House Cambridge University Engineering Department, CBL Seminar room BE4-38. For directions see http://learning.eng.cam.ac.uk/Public/Directions.

ClockWednesday 06 December 2017, 17:00-18:30

Reinforcement learning with a corrupted reward function

UserTom McGrath, Imperial College London.

House Cambridge University Engineering Department, CBL Seminar room BE4-38. For directions see http://learning.eng.cam.ac.uk/Public/Directions.

ClockWednesday 29 November 2017, 17:00-18:30

Solomonoff Induction and a Definition of Intelligence

UserJames Bell, Richard Ngo (University of Cambridge).

House Cambridge University Engineering Department, CBL Seminar room BE4-38. For directions see http://learning.eng.cam.ac.uk/Public/Directions.

ClockWednesday 22 November 2017, 17:00-18:30

Deep Reinforcement Learning from Human Preferences

UserJessica Yung (University of Cambridge).

House Cambridge University Engineering Department, CBL Seminar room BE4-38. For directions see http://learning.eng.cam.ac.uk/Public/Directions.

ClockWednesday 15 November 2017, 17:00-18:30

An introduction to adversarial attacks and defences

UserYingzhen Li (University of Cambridge).

House Cambridge University Engineering Department, CBL Seminar room BE4-38. For directions see http://learning.eng.cam.ac.uk/Public/Directions.

ClockWednesday 08 November 2017, 17:00-18:30

'Off-Switch Games' and Corrigibility

UserRichard Ngo (University of Cambridge).

House Cambridge University Engineering Department, CBL Seminar room BE4-38. For directions see http://learning.eng.cam.ac.uk/Public/Directions.

ClockWednesday 01 November 2017, 17:00-18:30

Cooperative Inverse Reinforcement Learning

UserRobert Pinsler (University of Cambridge).

House Cambridge University Engineering Department, CBL Seminar room BE4-38. For directions see http://learning.eng.cam.ac.uk/Public/Directions.

ClockWednesday 25 October 2017, 17:00-18:30

Engineering Safe AI seminar group

UserBeth Barnes, Computer Lab, University of Cambridge.

House Cambridge University Engineering Department, CBL Seminar room BE4-38. For directions see http://learning.eng.cam.ac.uk/Public/Directions.

ClockWednesday 18 October 2017, 17:00-18:30

Please see above for contact details for this list.

Other lists

AnneETyner LT0 https://us06web.zoom.us/j/81119667575?pwd=bXNQVzBXR2p3NXRDWDVVamFQa2JJdz09

Other talks

Loke CTR Annual Meeting 2026 - Epigenetics of embryogenesis and placentation Title TBC What Can a Spectrum Learn? AI Perspectives on the Kohn Laplacian of Sphere Quotients Tea and Coffee Break Title tbc Title to be confirmed
 

Β© 2006-2025 Talks.cam, University of Cambridge. Contact Us | Help and Documentation | Privacy and Publicity