The Artificial Intelligence In Science And The Science In Artificial Intelligence

Prof. Dr. Max Welling is a research chair in Machine Learning at the University of Amsterdam and a Distinguished Scientist at Microsoft Research. He is a fellow at the Canadian Institute for Advanced Research (CIFAR) and the European Lab for Learning and Intelligent Systems (ELLIS) where he also serves on the founding board. His previous appointments include VP at Qualcomm Technologies, professor at UC Irvine, postdoc at U. Toronto and UCL under supervision of prof. Geoffrey Hinton, and postdoc at Caltech under supervision of prof. Pietro Perona. He finished his PhD in theoretical high energy physics under supervision of Nobel laureate prof. Gerard ‘t Hooft.
Max Welling has served as associate editor in chief of IEEE TPAMI from 2011-2015, he serves on the advisory board of the Neurips foundation since 2015 and has been program chair and general chair of Neurips in 2013 and 2014 respectively. He was also program chair of AISTATS in 2009 and ECCV in 2016 and general chair of MIDL 2018. Max Welling is recipient of the ECCV Koenderink Prize in 2010 and the ICML Test of Time award in 2021. He directs the Amsterdam Machine Learning Lab (AMLAB) and co-directs the Qualcomm-UvA deep learning lab (QUVA) and the Bosch-UvA Deep Learning lab (DELTA).
What are the big recent advances in artificial intelligence (AI)? How can modern AI be effectively used in the natural sciences? Is there a deep relation between AI and physics? What societal challenges can we tackle with modern AI? What will the future of AI look like?
AI has progressed at an incredibly fast, sometimes scary, pace. From solving complicated games such as chess and GO, to generating images with uncanny quality, to Chatbots which exhibit the first signs of Artificial General Intelligence. Important as these advances are in their respective application domains, Max Welling predicts that we will see a much bigger paradigm shift in the way we will do scientific research in the near future based on these same techniques. For example, researchers at Google Deepmind used the same reinforcement learning technology that powers AlphaGO to fold proteins, cracking an open problem in the life sciences. Researchers around the world are making great strides in using the same technology that generates images to generate new drug molecule candidates. And the Large Language Models that power Chatbots like ChatGPT can be used to reason about complicated scientific problems.
These and future techniques can be used effectively to tackle some of the most urgent challenges that humanity will face over the next decades, such as designing new materials to capture and store carbon, making more efficient batteries, new catalysts to generate hydrogen more efficiently from water, and better designs for nuclear fusion reactors and wind turbines. But what is interesting is that the tools developed in the sciences are also at the core of some of the recent successes of AI. For instance, the models used to generate images are based on models developed by mathematicians and physicists a long time ago to describe liquids and gases. And with the advance of quantum computing, we may see yet another disruption at the interface of computer science and physics.
Max Welling will try to explain the enormous potential that this synergy will bring us, not only in terms of discovering powerful new AI algorithms, but also in terms of empowering the scientific applications that are relevant to tackle the big challenges for humanity over the decades.
Event held in English.
For more science visit:
• Website: www.scienceandcocktails.org
• Facebook: / scienceandcocktailscph
• UA-cam: / sciencecocktails
• Instagram: / scienceandcocktailsglobal
For more cool science visit:
• Website: www.scienceandcocktails.org
• Facebook: / scienceandcocktailscph
• UA-cam: / sciencecocktails
• Instagram: / scienceandcocktailsglobal

[@petersq5532]

31:00: transformer models. while NLP models works on conditional probability when you scale up the model it shows a sign if accured traits.
that is surprising and scary as at some point the accused trait could be intelligence and sentinel hypothetically.
(accured trait: a new trait appearing from complexity which was not part of the original design or the sumof the parts. like chemistry -> biology)

[@corinthiabowman5299]

Did anyone else see a wine glass with a kidney bean on top instead of a chair? 😂 I tried really hard to see the chair after he said it was a chair and my brain said "nope... wine glass and kidney bean"
🤣🤣

[@petersq5532]

33:20: thinking out of the box. there is no box for AI. box is for humans where they limited to their experience, intellect and creativity. math has no box so the resulting equatation optimalizasion is not bound to human constrains

[@petersq5532]

17:53: chair. your answer is limited to your experience, knowledge and creativity. for me the pic more like a bent red cell, erytrocyte.
machine learning: finding parameters to a mathematical equatation.
neural network: several equatation interact (calculation results passed to another equatation). learning is to find the parameters for each equatations.
deep learning: more equatations interonnected inloops to find the parameters.
there is no intelligence or understanding behind the result. th are is math, and constants and random numbers. sometimes thousands of them.
ll the rest is mystifying and bluring the plain reality. iseing borrowed terminology from neurobiology helps to mislead the public. the social implication if specialised AI tools nevertheless not less serious