Slonim speaks at JURIX 2014
We are pleased to inform that one of the invited speakers at JURIX 2014 is Noam Slonim (PhD), Senior Research Staff Member at the Analytics Department at IBM.
The anticipated topic of Noam’s talk is: “IBM debating technologies”.
After that, Noam spent four years as a postdoc at the Department of Physics at Princeton University, working in the Lewis-Sigler Institute for Integrative Genomics, mainly on the development and application of novel techniques to the analysis of genomics data.
After his postdoc Noam joined the IBM Haifa Research Lab at 2007, as a Research Staff Member, and since then led various Machine Learning and Text analytics projects, that were recognized by several IBM awards, including an IBM Outstanding Technical Achievement Award.
Since 2011, Noam serves as the IBM Research Technical Lead of the IBM Grand Challenge, entitled “IBM Debating Technologies”.
Prof. dr Pieter Adriaans speaks at JURIX 2014
Facticity and meaningful information
Pieter Adriaans, University of Amsterdam
There is a growing interest in theories and techniques that allow us to measure the complexity of legal databases and information systems (Katz and Bommarito II, Jurix 2011). The theory of facticity, developed in our group in Amsterdam, studies the interaction between structured and non-structured information in (large) data sets and it might be relevant in this context. A phonebook contains a lot of information, but its structure is not very complex. A legal database or a description of the brain (cell/wheather/society etc.) will not only contain a lot of information, but will be inherently complex due to the high number of interrelations. This difference between information and model complexity is not captured by the current theories. All well-known quantitative information measures (specifically Shannon Information and Kolmogorov complexity) assign the highest information content to data sets with the highest entropy. In this sense a television broadcast with only white noise would contain the most meaningful information. This is counterintuitive. Currently we are studying the problem of the balance between model information and non-structured data in a database, in terms of two-part code optimization, under the unifying perspective of Kolmogorov complexity. This study reveals the existence ‘of a landscape of measures of meaningful information’ with interesting phase transitions and non-linear phenomena. These non-entropy based measures are relevant for the characterization of large knowledge graphs since they capture the notion of the amount of useful information in a data set. In my lecture I will explain the theoretical backgrounds of this work and discuss practical methods to apply these measures to large knowledge bases.
Short Biographical note
Over the years Pieter Adriaans (1955) has built up an impressive unusually broad oeuvre that varies from paintings and sculptures to installations, books, scientific papers, patents and musical compositions. This achievement is remarkable given the fact that Adriaans also has a master in philosophy, a PhD in theoretical computer science and has, together with his business partner Dolf Zantinge, founded a very successful computer company. In 1992 got his doctorate at the university of Amsterdam and in 1998 he was appointed professor of learning and adaptive systems at the same institute. He is also fellow of the Info-Metrics Institute at the American University in Washington. His main research interests are:
Pieter and his wife Rini live in Kockengen in the Netherlands and part of the year on the island of São Jorge, one of the Azores.