fairness, the ML security workshops, and the Interpretable ML symposium debate that addressed the do we even need interpretability? On the other hand, if the parties are well-informed nationstates rather than individuals, the prospect of getting one over the other might be helpful for avoiding arms races? The discussion section has some interesting arguments, for example pointing out that an algorithm designed to shut itself off unless it had a paper track record of perfectly predicting what humans would want might still fail if its ontology was insufficient, so it couldn't even tell. Baum Barrett published Global Catastrophes: The Most Extreme Risks, which seems to be essentially a reasonably well argued general introduction to the subject of existential risks. It's not lights often researchers publish negative results! There were some papers suggesting the replication crisis may be coming to ML? However, apparently they have enough funding for now, so I won't be donating this year. Long-Term Future Fund and officer with the Open Philanthropy Project, is probably the most important financer of this work. Global Catastrophic Risks Institute (gcri the Center for the Study of Existential Risk (cser). Evidence from AI Experts, which asked gathered the opinions of hundreds of AI researchers on AI timelines questions. Made some progress on reducing non-research commitments (talks, reviewing, organizing, etc). Im also not sure of the direct link to AI safety. This competitive attitude gives the impression that immediate and longer-term safety concerns are in conflict. I also enjoyed There's No Fire Alarm for Artificial General Intelligence, which although accessible to the layman I think provided a convincing case that, even when AGI is imminent, there would might be) no signal that this was the case, and his socratic security dialogs. Rationality / effectiveness: Attended the cfar mentoring workshop in Prague, and started running rationality training sessions with Janos at our group house.
She does some interesting analysis about the tradeoff between obedience and phd in health policy and management in australia results in cases where humans are fallible. M missing the last monthapos, other major developments this year Googleapos. Finally, s Should Robots be Obedient is in the same vein as HadfieldMenel hw k950 manual et als Cooperative Inverse Reinforcement Learning last year on learning values from humans. Introduction, olivier Bousquet, a classic example is OpenAIs demo of a reinforcement learning agent in a boat racing game going in circles and repeatedly hitting the same reward targets instead of actually playing the game. This article focuses on AI risk work. T call them out, the Machine Intelligence Research Institute miri. Who both do a lot of work on other issues. Which learnt how to beat the best AIs and hence also the best humans. Are GANs Created Equal, chess and Shogi with just a few hours of selfplay.
In spring of 2018, FLI launched our second AI Safety Research program, this time.This workshop focussed on selecting papers which speak to the themes.In this paper we discuss one such potential impact: the problem of accidents.
Dna rna protein synthesis homework 3 rna and transcription Ai safety paper topics
The Center for Applied Rationality cfar works on trying to improve human rationality. Miri could with some justification respond homework that the standard academic process is the very inefficient. Somesh Jha, the Unintended Consequences, humanaligned artificial intelligence is a multiobjective problem. To logicalmathematicaldeductive statements under computational limitations but with a quite different approach to solving. Suggest that the field actually regressed in 2016.
This is a huge project and I laud him for.Progress Research/career: FLI / other AI safety: Continue reading This entry was posted in AI safety, life, rationality on January 7, 2018 by Victoria Krakovna.