Timo Honkela: Artificial Intelligence and Machine Learning in the Service of the Good
2 likes1,172 views
The document discusses artificial intelligence and machine learning and their potential applications. It notes that AI is being used in many services from large tech companies and was originally developed in universities. It also discusses different AI paradigms and gives examples of research from the University of Helsinki applying machine learning to tasks like pattern recognition, sentiment analysis, machine translation and more. The document expresses optimism that AI can be developed and applied to benefit society by addressing challenges in healthcare, transportation, education and more, while also needing to consider ethical concerns.
Timo Honkela: Artificial Intelligence and Machine Learning in the Service of the Good
- 1. TIMO HONKELA: AI & ML IN THE SERVICE OF THE GOOD 2 Nov 2015 Timo Honkela AI & ML IN THE SERVICE OF THE GOOD timo.honkela@helsinki.fi
- 2. TIMO HONKELA: AI & ML IN THE SERVICE OF THE GOOD Anticipation, prediction, time series analysis Some human skills being simulated Pattern recognition, machine vision, speech processing Natural language understanding and generation Symbolic, probabilistic and intuitive reasoning Problem solving and creativity Emotional presence and social activity
- 3. TIMO HONKELA: AI & ML IN THE SERVICE OF THE GOOD Why a new situation? ● Data, data, bid data – data-driven approaches that enable going even far beyond the understanding of a single person or even of a traditional human research community ● Machine-learning-based artificial intelligent systems have become core modules in many services provided by companies like Google, IBM, Facebook, Amazon, etc. ● The majority of the original ideas and innovations stem from university research, though (whether we consider IBM Watson, Google Translate, etc.)
- 4. TIMO HONKELA: AI & ML IN THE SERVICE OF THE GOOD AI paradigms dichotomized ● Theory-driven versus data-driven ● Deductive versus inductive ● Logical reasoning versus statistical/ probabilistic/neural reasoning ● Modeling explicit rational thinking versus intuitive and implicit thinking and problem solving ● Symbolic representation versus pattern/vector representations ● Ontologies versus word and document vector spaces
- 5. TIMO HONKELA: AI & ML IN THE SERVICE OF THE GOOD Pattern recognition Kurimo et al. Speech recognition http://www.cis.hut.fi/cis/projects/speech/
- 6. TIMO HONKELA: AI & ML IN THE SERVICE OF THE GOOD Pattern classification Vuori & Laaksonen 2002 Recognition of handwritten characters
- 7. TIMO HONKELA: AI & ML IN THE SERVICE OF THE GOOD Pattern recognition Koskela, Sjöberg & Laaksonen 2007 Features clustering
- 8. TIMO HONKELA: AI & ML IN THE SERVICE OF THE GOOD Pattern Generation http://googleresearch.blogspot.fi/2015/06/inceptionism-going-deeper-into-neural.html
- 9. TIMO HONKELA: AI & ML IN THE SERVICE OF THE GOOD Building and using vast corpora http://www.csc.fi/kielipankki/https://korp.csc.fi K. Lindén et al.
- 10. TIMO HONKELA: AI & ML IN THE SERVICE OF THE GOOD Classical example: Learning meaning from context: Maps of words in Grimm fairy tales Honkela, Pulkki & Kohonen 1995 Predates topic modeling
- 11. TIMO HONKELA: AI & ML IN THE SERVICE OF THE GOOD Multimodally Grounded Language Technology A project funded by Academy of Finland 2011-2014 Klaus Förfer's thesis: https://aaltodoc.aalto.fi /handle/123456789/17857 http://cogsys.blogspot.fi/2015/10/ klaus-forger-from-motion-capture-to.html
- 12. TIMO HONKELA: AI & ML IN THE SERVICE OF THE GOOD Honkela, Korhonen, Lagus & Saarinen: Five-dimensional sentiment analysis of corpora, documents and words, Proceedings of WSOM 2014 Sentiment analysis
- 13. TIMO HONKELA: AI & ML IN THE SERVICE OF THE GOOD Machine translation http://www.aclweb.org/anthology/W/W15/W15-18.pdf#page=219 Tiedemann 2015 Aligning linguistics expressions
- 14. TIMO HONKELA: AI & ML IN THE SERVICE OF THE GOOD Timo Honkela AI & ML IN THE SERVICE OF THE GOOD IN THE SERVICE OF THE GOOD IN THE SERVICE OF THE GOOD 2 Nov 2015
- 15. TIMO HONKELA: AI & ML IN THE SERVICE OF THE GOOD Life saving systems
- 16. TIMO HONKELA: AI & ML IN THE SERVICE OF THE GOOD Robotic cars could save one million lives each year? (*) “Tesla CEO Elon Musk referred to each Model S owners as an “expert trainer” – meaning that each driver will train the autonomous features of the system to feed the collective network intelligence of the fleet by simply driving the electric vehicle on Autopilot.” (*) https://en.wikipedia.org/wiki/List_of_countries_by_traffic-related_death_rate http://electrek.co/2015/10/30/the-autopilot-is-learning-fast-model-s-owners-are-already-reporting-that-teslas-autopilot-is-self-improving/
- 17. TIMO HONKELA: AI & ML IN THE SERVICE OF THE GOOD An example: Resolving Transferin Cancer Classification PR Adhikari & J Hollmén Discovery Science, 1-8 Jaakko Hollmén's talk New approaches to diagnosis and treatment
- 18. TIMO HONKELA: AI & ML IN THE SERVICE OF THE GOOD Towards Peace and Prosperity
- 19. TIMO HONKELA: AI & ML IN THE SERVICE OF THE GOOD Jörg Tiedemann's talk Building bridges over language boundaries
- 20. TIMO HONKELA: AI & ML IN THE SERVICE OF THE GOOD Understanding subjectivity and contextuality of human knowledge Grounded Intersubjecity Concept Analysis Towards Within-Language TranslationTranslation
- 21. TIMO HONKELA: AI & ML IN THE SERVICE OF THE GOOD Understanding historical developments through the analysis of large historical corpora difi.kansalliskirjasto.fi OCR problems? An AI & ML challenge!
- 22. TIMO HONKELA: AI & ML IN THE SERVICE OF THE GOOD Understanding people and the society through the analysis of large social media corpora such as Suomi24 Krista Lagus' talk
- 23. TIMO HONKELA: AI & ML IN THE SERVICE OF THE GOOD Increased understanding of complex socio-economical systems through crowdsourcing, data analysis and simulation
- 24. TIMO HONKELA: AI & ML IN THE SERVICE OF THE GOOD Information Society for All
- 25. TIMO HONKELA: AI & ML IN THE SERVICE OF THE GOOD Liisa Tiittula & Mikko Kurimo' talk Jorma Laaksonen's talk
- 26. TIMO HONKELA: AI & ML IN THE SERVICE OF THE GOOD Conclusions
- 27. TIMO HONKELA: AI & ML IN THE SERVICE OF THE GOOD ● Many positive developments are taking place and will be possible ● Negative effects and scenarios have to be taken into account, too ● Interdisciplinary theoretical and practical work on philosophical questions, ethical concerns, societal impacts (privacy, work), economics (fairness), complex themes related to humanities, etc. is necessary ● In general, I am optimistic that the world can be made to be a better place by appropriate and wise use of artificial intelligence and machine learning
- 28. TIMO HONKELA: AI & ML IN THE SERVICE OF THE GOOD P.S. Recently professor Bengt Holmström reminded that at MIT studies in humanities and social sciences are obligatory for all students The “two cultures” to work together for the best for all
- 29. TIMO HONKELA: AI & ML IN THE SERVICE OF THE GOOD Thank you for your attention!
- 30. TIMO HONKELA: AI & ML IN THE SERVICE OF THE GOOD Discussion forum: https://goo.gl/ixgW8c