- SIPTA News - lists.sipta.org

[SIPTA] Fwd: [UAI] CfP: Reliable Machine Learning in the Wild at ICML 2017, deadline 17 July
by Sébastien Destercke July 11, 2017

July 11, 2017

Dear all, This might interest some of you, in particular the robustness aspects first advocated. ---------- Forwarded message ---------- From: Adrian Weller <aw665(a)cam.ac.uk> Date: 2017-07-10 0:48 GMT+02:00 Subject: [UAI] CfP: Reliable Machine Learning in the Wild at ICML 2017, deadline 17 July To: Cc : Jacob Steinhardt <jacob.steinhardt(a)gmail.com>, Dylan Hadfield-Menell <dhm(a)eecs.berkeley.edu>, Smitha Milli <smilli(a)berkeley.edu> Final call for Papers for ICML 2017 Workshop on Reliable Machine Learning in the Wild, please forward to others who may have interest. Workshop website https://sites.google.com/site/wildml2017icml/ When can we trust that a system that has performed well in the past will continue to do so in the future? Designing systems that are reliable in the wild is essential for high stakes applications such as self-driving cars and automated surgical assistants. This workshop aims to bring together researchers in diverse areas such as reinforcement learning, human-robot interaction, game theory, cognitive science, and security to further the field of reliability in machine learning. We will focus on three aspects — robustness (to adversaries, distributional shift, model misspecification, corrupted data); awareness (of when a change has occurred, when the model might be miscalibrated, etc.);and adaptation (to new situations or objectives). We aim to consider each of these in the context of the complex human factors that impact the successful application or meaningful monitoring of any artificial intelligence technology. Together, these will aid us in designing and deploying reliable machine learning systems. We are seeking submissions that deal with the challenges of reliably applied machine learning techniques in the real world. Some possible questions touching on each of these categories are given below, though we also welcome submissions that do not directly fit into these categories. Robustness: How can we make a system robust to novel or potentially adversarial inputs? What are ways of handling model mis-specification or corrupted training data? What can be done if the training data is potentially a function of system behavior or of other agents in the environment (e.g. when collecting data on users that respond to changes in the system and might also behave strategically)? Awareness: How do we make a system aware of its environment and of its own limitations, so that it can recognize and signal when it is no longer able to make reliable predictions or decisions? Can it successfully identify “strange” inputs or situations and take appropriately conservative actions? How can it detect when changes in the environment have occurred that require re-training? How can it detect that its model might be mis-specified or poorly-calibrated? Adaptation: How can machine learning systems detect and adapt to changes in their environment, especially large changes (e.g. low overlap between train and test distributions, poor initial model assumptions, or shifts in the underlying prediction function)? How should an autonomous agent act when confronting radically new contexts? Monitoring: How can we monitor large-scale systems in order to judge if they are performing well? If things go wrong, what tools can help? Value Alignment: For systems with complex desiderata, how can we learn a value function that captures and balances all relevant considerations? How should a system act given uncertainty about its value function? Can we make sure that a system reflects the values of the humans who use it? Reward Hacking: How can we ensure that the objective of a system is immune to reward hacking? Reward hacking is a way that the system can attain high reward that was unintended by the system designer. For example see https://blog.openai.com/faulty-reward-functions/ Human Factors: Actual humans will be interacting and adapting to these systems when they are deployed. How do properties of humans affect the guarantees of performance that the system has? What if the humans are suboptimal or even adversarial? How to submit Papers submitted to the workshop should be up to four pages long excluding references and in ICML 2017 format. They should be submitted via Easychair at https://easychair.org/conferences/?conf=rmlw17 . As the review process is not blind, authors can reveal their identity in their submissions. Accepted submissions will be presented as posters or talks. We will accept submissions at two deadlines. One earlier deadline, with an earlier acceptance notification, and one later one. Our goal is to allow for late submission to the extent that we can, while still allowing some people to get early confirmation of paper acceptance, which they might need in order to arrange travel in time. Important Dates: Submission deadline 1: 16 June 2017 Acceptance notification 1: 1 July 2017 Submission deadline 2: 17 July 2017 Acceptance notification 2: 31 July 2017 Final camera-ready versions of accepted papers: 5 August 2017 Workshop: 11 August 2017 Thank you, Dylan, Jacob, Smitha and Adrian ---------------------------------------------- Adrian Weller _______________________________________________ uai mailing list uai(a)ENGR.ORST.EDU https://secure.engr.oregonstate.edu/mailman/listinfo/uai -- ==================================== Sebastien Destercke, Ph. D. CNRS researcher in computer science. Université de Technologie de Compiegne U.M.R. C.N.R.S. 7253 Heudiasyc Avenue de Landshut F-60205 Compiegne Cedex FRANCE Tel: +33 (0)3 44 23 79 85 Fax: +33 (0)3 44 23 44 77 ====================================

1 0

[SIPTA] Fwd: [Bayes-forum: 148] Three post-doctoral positions on Bayesian machine learning
by Thomas Augustin June 27, 2017

June 27, 2017

1 0

Re: [SIPTA] Question about the word "gamble" -- and "prevision"
by Serena Doria June 12, 2017

June 12, 2017

Dear Professor Bickis, In Italian the two words Previsione=Prevision and Predizione=Prediction have different acceptation. (Devoto-Oli Dictionary) Previsione = Elaborazione mentale relativa al futuro, sulla base di indizi piu' o meno sicuri (Mental thinking about the future based on more or less secure clues) Example: Previsioni atmosferiche =Weather forecasts Predizione = previsione autorevole e solenne riferita come oggetto di una previsione o profezia (Authoritative and solemn predictions referred as the subject of a prediction or prophecy) Best regards, Serena Professor Bickis <bickis(a)snoopy.usask.ca> ha scritto: > Thanks for all your collective erudition. I am aware that “gamble” > or its synonyms “wager” and “bet” have been used by many writers in > the context of probability and decisions theory. I was thinking of > the use of the word in the generic abstract sense that Walley uses. > Since Walley’s development picks up on de Finetti and Williams in > using “prevision”, I wondered if one of them had also used “gamble”, > but it seems that they did not. > > Googling some Italian web sites on probability, I did come across > the word “scommessa” in discussion of expectation/prevision, but I > wonder if de Finetti himself used it in the sense that Walley used > “gamble”. > > I am also a bit mystified by the history of the word “prevision”. > Writing in French his 1937 paper, de Finetti uses “prévision” rather > than the more common (even then) espérance. I would translate > “prévision” as “predicition” in English. I see that in the 1974 > English translation of de Finetti’s "Teoria de la Probabilità”, the > translators note that they are using “prevision” (where earlier > translators had used “foresight”), thereby making “prevision” less > esoteric in English. However, in the same book, de Finetti makes a > big deal that the “prevision is not prediction” (in the English > translation). I would be interested if any Italian speakers out > there could explain how de Finetti described this in his native > language. I would render the French “prévision” as “predicition” in > English, so naively would think that “previsione” in Italian also > translates as “prediction”. Google Translate, however, gives me > “predizione”, > for “predicition”. So when writing in Italian does de Finetti > contrast “previsione” from “predizione”? Here once can see the > difference between seeing and speaking, which is somewhat masked in > English. Is that what de Finetti was driving at? > > Interested in discussion. > > Mik Bickis > >> On Jun 6, 2017, at 08:18 AM, Samuel Cohen <cohens(a)maths.ox.ac.uk> wrote: >> >> Dear all, >> >> If you are looking more generally at the history, it's unsurprising >> the word 'gamble' wasn't used much before the mid-20th century. >> While it was a common word in english since the 18th century (but >> was a colloquialism), it seems to have moved up in popularity post >> 1900 (see for example google ngrams). >> >> De Moivre, in arguably the first textbook on probability in >> English, uses 'wager' throughout (which is natural, as his >> attention is on games of chance). >> >> Best, >> Sam Cohen >> >> >> >> so On 06/06/17 14:01, Teddy Seidenfeld wrote: >>> Dear Mik, Erik, and Friends, >>> >>> I'm no scholar of the term 'gamble', but it's been used by many >>> decision theorists in closely related senses. >>> >>> For instance, Savage's (1954) section 5.2 (titled 'Gambles') uses >>> it within his theory (of subjective expected utility) to refer to >>> the equivalence class of simple acts that carry the same >>> (personal) probability distribution over consequences. Savage is >>> generalizing the von Neumann-Morgenstern sense of 'gamble'. >>> >>> I grew up with the von Neumann-Morgenstern's theory of cardinal >>> utility for gambles -- though vN-M use an extraneous concept of >>> 'probability' in contrast with Savage's approach. >>> >>> Best, >>> Teddy >>> >>> >>> >>> >>> >>> >>> On Mon, Jun 5, 2017 at 7:12 PM, Professor Bickis >>> <bickis(a)snoopy.usask.ca> wrote: >>> Hello: >>> >>> Can someone answer this question? >>> >>> Was it Walley who initiated the word “gamble” to use in place of >>> the classical “random variable”? De Finetti uses “random >>> quantity”, as does Williams. “Gamble” is ubiquitous in the >>> current literature, but I have not seen it anywhere prior to >>> Walley’s 1991 book. >>> Walley does not seem to attribute the term to anyone else. >>> >>> Mik Bickis >>> _______________________________________________ >>> SIPTA mailing list >>> SIPTA(a)idsia.ch >>> http://mailman2.ti-edu.ch/mailman/listinfo/sipta >>> >>> >>> >>> _______________________________________________ >>> SIPTA mailing list >>> >>> SIPTA(a)idsia.ch >>> http://mailman2.ti-edu.ch/mailman/listinfo/sipta >> >> _______________________________________________ >> SIPTA mailing list >> SIPTA(a)idsia.ch >> http://mailman2.ti-edu.ch/mailman/listinfo/sipta > > _______________________________________________ > SIPTA mailing list > SIPTA(a)idsia.ch > http://mailman2.ti-edu.ch/mailman/listinfo/sipta >

3 2

Re: [SIPTA] Question about the word "gamble"
by Samuel Cohen June 7, 2017

June 7, 2017

3 2

[SIPTA] Question about the word "gamble"
by Professor Bickis June 6, 2017

June 6, 2017

Hello: Can someone answer this question? Was it Walley who initiated the word “gamble” to use in place of the classical “random variable”? De Finetti uses “random quantity”, as does Williams. “Gamble” is ubiquitous in the current literature, but I have not seen it anywhere prior to Walley’s 1991 book. Walley does not seem to attribute the term to anyone else. Mik Bickis

3 2

[SIPTA] Fw: IPMU 2018 - Mark your calendars
by Kreinovich June 4, 2017

June 4, 2017

1 0

[SIPTA] Fwd: [ST-NET] 9 Independent Research Group Leaders Tuebingen/Stuttgart (Germany)
by Thomas Augustin June 2, 2017

June 2, 2017

1 0

[SIPTA] [UAI] IUKM 2018--CfP: March 15 - 17, 2018, Ha Noi, Vietnam
by Van-Nam HUYNH June 2, 2017

June 2, 2017

[Apologies for cross-posting. Your help in circulating this call for papers to potentially interested parties will be highly appreciated.] ————————————————————————————————————— IUKM 2018 — CALL FOR PAPERS ————————————————————————————————————— The Sixth International Symposium on INTEGRATED UNCERTAINTY in KNOWLEDGE MODELLING and DECISION MAKING http://www.jaist.ac.jp/IUKM/IUKM2018/ March 15 - 17, 2018, Hanoi, VIETNAM ————————————————————————————————————— -The International Symposium on Integrated Uncertainty in Knowledge Modelling and Decision Making (IUKM) aims to provide a forum for exchanges of research results, ideas for and experience of application among researchers and practitioners involved with all aspects of uncertainty management and application. Previous symposia have been successfully held in Ishikawa, Japan (IUM 2010 - AISC 68), Hangzhou, China (IUKM 2011 - LNAI 7027), Beijing, China (IUKM 2013 - LNAI 8032), Nha Trang, Vietnam (IUKM 2015 - LNAI 9376), and Da Nang, Vietnam (IUKM 2016 - LNAI 9978). The sixth IUKM (2018) will take place in the capital of Hanoi, Vietnam, during 15-17 March, 2018, and be jointly organized by Hanoi National University of Education (HNUE), Japan Advanced Institute of Science and Technology (JAIST), Belief Functions and Applications Society (BFAS). ————————————————————————————————————— Proposal for special sessions and workshops is encouraged and welcome. ————————————————————————————————————— Relevant topics for the IUKM 2018 Symposium include, but are not limited to, the following: + Theory and Methodology - Uncertainty formalisms: Bayesian probability, Dempster-Shafer theory, imprecise probability, random sets, rough sets, fuzzy sets and interval-based models - Uncertainty modeling and management in big-data environments - Learning and reasoning with uncertainty - Information fusion and knowledge integration in uncertain environments - Decision making under various types of uncertainty - Aggregation operators for decision making - Copulas for dependence modelling - Granular and soft computing - Computational intelligence + Application - Data mining and knowledge discovery - Ontology engineering and Semantic Web - Intelligent data analysis and modelling - Agents and argumentation - Natural language processing - Medical informatics and bioinformatics - Ranking and recommendation systems - Big data and cloud computing - Social network analysis and mining - Sensor fusion - System identification and modelling - Diagnosis and reliability - Decision support systems - Kansei/affective engineering - Service computing - Engineering management - Supply chain management - Environmental management - Economics and econometrics PUBLICATION The authors are invited to submit their full papers by 15 Sept. 2017 (extended deadline). The submissions will be peer-reviewed for originality and scientific quality. As done in previous editions, the proceedings of IUKM 2018 will be published by Springer-Verlag in the Lecture Notes in Artificial Intelligence series, and be available at the Symposium. Authors are requested to prepare their papers in the Springer LNAI format. Submissions should not exceed 12 pages and must be submitted electronically through the conference’s Easychair submission page (https://easychair.org/conferences/?conf=iukm2018) IMPORTANT DATES Deadline for paper submission: 15 Sept. 2017 Notification of acceptance/rejection: 20 Nov., 2017 Final camera-ready and author registration: 15 Dec., 2017 Early non-author registration: 10 Jan. 2018 Symposium dates: 15-17 March, 2018 —————————————

1 0

[SIPTA] IP question on math.stackexchange
by Abrams May 29, 2017

May 29, 2017

1 0

[SIPTA] Positions (2 PhD + 1 post-doctoral) at Compiègne University, France
by Sébastien Destercke May 10, 2017

May 10, 2017

Dear all, The Heudiasyc Laboratory is looking for motivated candidates on different positions (2 PhD + 1 post-doctoral) with funded scholarships, all concerning the handling of severe uncertainties and all starting around september/october 2017: 1. PhD position on machine learning with structured outputs: "Learning and predicting graphs from limited information". Details in: https://www.hds.utc.fr/~sdesterc/dokuwiki/_media/thesis_graphs.pdf 2. PhD position on inverse problems, with focus on non-linear mechanical models: "Solving the inverse problem under severe uncertainty and with multiple information sources" Details in: https://www.hds.utc.fr/~sdesterc/dokuwiki/_media/thesis_meca.pdf 3. Post-doctoral position on designing reliable systems: "Designing safe systems under severe uncertainties" Details in: https://www.hds.utc.fr/~sdesterc/dokuwiki/_media/postdoc_reliability.pdf If interested, you can contact me by mail. Best regards Sebastien Destercke P.S.: please distribute this announcement to potentially interested and interesting candidates. -- ==================================== Sebastien Destercke, Ph. D. CNRS researcher in computer science. Université de Technologie de Compiegne U.M.R. C.N.R.S. 7253 Heudiasyc Avenue de Landshut F-60205 Compiegne Cedex FRANCE Tel: +33 (0)3 44 23 79 85 Fax: +33 (0)3 44 23 44 77 ====================================

1 0