Events: Data- och informationsteknik events at Chalmers University of TechnologyFri, 20 Apr 2018 15:58:42 +0200 Sjösten, Computer Science and Engineering<p>EA, lecture hall, EDIT trappa C, D och H, EDIT</p><p>​Guarding the Boundary: Information Flow Tracking in the Presence of Libraries</p>In modern software development, the use of libraries is prevalent. Libraries pose a big security challenge. How can we ensure that sensitive data is not being leaked through libraries? This is the first question of the thesis. We propose the use of information-flow control, by developing a principled approach for allowing information-flow tracking in libraries, even if they are written in a language not supporting information-flow control. With this approach, we allow for library functions to have unlabel and relabel models, explaining how values are unlabeled and relabeled when being marshaled between the labeled program and the library. These models are used in combination with lazy marshaling to handle structured data such as lists and records, higher-order functions and references.<br /><p></p> Modern browsers allow for browser modifications through browser extensions, which have special privileges and can, e.g., modify the DOM. As extensions can be intrusive, it is in a webpage's interest to know which extensions are installed in a browser. The second question of the thesis is if it is possible for a webpage to know which extensions are installed in the browser? We conduct a large-scale study to determine how many extensions that are detectable from a webpage based on the extension's resources, showing over 50% of the top 1000 Chrome extensions can be detected, as well as how many of the Alexa top 100,000 webpages employ the technique of the paper. Automatic Learning and Enforcement of Authorization Rules in Online Social Networks<p>EC, lecture hall, EDIT trappa C, D och H, EDIT</p><p>Yiannis Papagiannis, Facebook London</p><br />Abstract: <br />In this talk we will discuss Invariant Detector (IVD), a defense-in-depth system that automatically learns authorization rules from normal data manipulation patterns and distills them into likely invariants. IVD acts as an additional layer of defense, working behind the scenes, complementary to privacy frameworks and testing. We have designed and implemented IVD to handle the unique challenges posed by modern online social networks. IVD is currently running at Facebook, where it infers and evaluates daily more than 200,000 invariants from a sample of roughly 500 million client requests, and checks the resulting invariants every second against millions of writes made to a graph database containing trillions of entities. <br /><br /><a href=""></a><br /><a href=""></a> Road seminar: Space and Satellite<p>RUAG Space, Solhusgatan 11, Göteborg</p><p>​Sweden and especially West Sweden have many world-leading space and satellite companies. Listen to a few of them at this Microwave Road Event. Welcome!</p>​ <br />Satellite communication is entering a revolutionary phase in bringing affordable broadband to every corner of the earth with new advanced microwave and antenna technology. New High Throughput Satellites (HTS) have 100 - 1000 times more capacity than traditional TV broadcast satellites. This will mean significantly lower costs for using broadband via satellites and open completely new business opportunities in large parts of the world where 4 billion people still lack broadband connectivity. Weather, remote sensing, GPS, and scientific satellites will continue to be important for society.  Sweden and especially West Sweden have many world-leading space and satellite companies. Listen to a few of them at this Microwave Road Event April 25th, 2018.<br /><br /><span><strong>Regster at the latest on 18th April 2018</strong>, to <a href=""></a>.<span style="display:inline-block"></span></span><br /><br />PROGRAM<br />16.00–16.30 Registration and Coffee<br />16.30–16.50 <strong>Deborah Lygonis</strong>, Innovatum, “ESA Business Incubation Centre in Sweden”<br />16.50–17.10 <strong>Mats Andersson</strong>, Forsway, ”Smart integration of satellite and terrestrial infrastructure”<br />17.10–17.30 <strong>Jörgen Nilsson</strong>, RUAG Space, “Q- and V-band converters for high throughput<br />communication satellites”<br />17.30–18.00 Coffee break and Scholarship ceremony<br />18.00–18.20 <strong>Jakob Kallmér</strong>, Satcube, “Lightweight terminal using high throughput satellites anywhere”<br />18.20–18.40 <strong>Anders Emrich</strong>, Omnisys, ”Microwave and THz instruments in space”<br />18.40–19.00 <strong>Oleg Lupikov</strong>, Chalmers, “Digital beamforming focal plane arrays for space-borne passive<br />ocean remote sensing”<br />19.00–21.00 Beverage, food and continuation space and satellite discussions models of text normalization for speech applications<p>Palmstedtsalen, Chalmers conference centre, Chalmersplatsen 1, Johanneberg</p><p>​Speaker: Richard Sproat, Research Scientist, Google Research, New York</p>​(Joint work with Ke Wu, Hao Zhang, Kyle Gorman, Felix Stahlberg, Xiaochang Peng and Brian Roark).<br /><br /><strong>Abstract:</strong> Speech applications such as text-to-speech (TTS) or automatic speech recognition (ASR), must not only know how to read ordinary words, but must also know how to read numbers, abbreviations, measure expressions, times, dates, and a whole range of other constructions that one frequently finds in written texts. The problem of dealing with such material is called text normalization. The traditional approach to this problem, and the one currently used in Google’s deployed TTS and ASR systems, involves large hand-constructed grammars, which are costly to develop and tricky to maintain. It would be nice if one could simply train a system from text paired with its verbalization.<br /><br />I will present our work on applying neural sequence-to-sequence RNN models to the problem of text normalization. Given sufficient training data, such models can achieve very high accuracy, but also tend to produce the occasional error — reading “kB” as “hectare”, misreading a long number such as “3,281” — that would be problematic in a real application. The most powerful method we have found to correct such errors is to use finite-state over-generating covering grammars at decoding time to guide the RNN away from “silly” readings: Such covering grammars can be learned from a very small amount of annotated data. The resulting system is thus a hybrid system, rather than a purely neural one, a purely neural approach being apparently impossible at present.<br /><br /><strong>Brief bio:</strong> Richard Sproat is a Research Scientist at Google Research, New York. From 2009-2012 he was a professor at the Center for Spoken Language Understanding at the Oregon Health and Science University. Prior to going to OHSU, he was a professor in the departments of Linguistics and Electrical and Computer Engineering at the University of Illinois at Urbana-Champaign. He was also a full-time faculty member at the Beckman Institute. He still holds adjunct positions in Linguistics and ECE at UIUC. <br />Before joining the faculty at UIUC he worked in the Information Systems and Analysis Research Department headed by Ken Church at AT&amp;T Labs --- Research where he worked on Speech and Text Data Mining: extracting potentially useful information from large speech or text databases using a combination of speech/NLP technology and data mining techniques.<br />Before joining Ken's department he worked in the Human/Computer Interaction Research Department headed by Candy Kamm. His most recent project in that department was WordsEye, an automatic text-to-scene conversion system. The WordsEye technology is now being developed at Semantic Light, LLC.<br /><a href=""></a><br /> Dolonius, Computer Science and Engineering<p>EL41, lecture room, Linsen, EDIT</p><p>Sparse Voxel DAGs for Shadows and for Geometry with Colors​</p> Liebel, Computer Science and Engineering<p>Beta, lecture hall, Saga, Campus Lindholmen</p><p>​An Empirical Investigation of Using Models During Requirements Engineering in the Automotive Industry</p>​Grischa Liebel belongs to the Software Engineering division of Computer Science and Engineering. <br /><br /><span class="text-normal"><span class="text-normal">Opponent<br />Professor Jon Whittle, Monash University, Australia. <br /><span class="text-normal"><br />Grading Committee<br />Tao Yue, Simula Research Laboratory, Norway. <br />Anna Perini, Fundazione Bruno Kessler, Italy. <br />Professor Juergen Dingel, Queens University and Royal Military College, Canada. </span></span></span><br /><br /> Symposium on Deep Learning 2018<p>Wallenberg Conference Center, Medicinaregatan 20, Gothenburg</p><p>​The 2nd Swedish Symposium on Deep Learning will be held at Chalmers University of Technology, 4-5 September, 2018.</p>​ <br />The premier event in Sweden bringing together top researchers in Deep Learning across academia and industry. This year there will be thematic sessions on Vision, Natural Language Technologies and Health Engineering, three of the most high impact areas for Deep Learning today.<br /><br /><strong>Keynote Speakers:</strong><br /><ul><li><a href="" target="_blank">Chris Dyer</a>, Carnegie Mellon University (confirmed) </li> <li>Trevor Darrell, U.C. Berkeley (to be confirmed) </li> <li>David Nistér, Vice-President, NVIDIA self-driving cars (Sweden) (to be confirmed) </li> <li><a href="" target="_blank">Hossein Azizpour</a>, KTH Royal Institute of Technology (confirmed)</li></ul>   <br /><span><p><br /></p> <p>The programme will be continuously updated and announced.</p></span> Sustainability Day 2018<p>TBA</p><p>​Chalmers Sustainability Day is back - Tuesday 23 October 2018. This year’s theme is Good Health and Well-being.</p>​ <br />More information will follow, stay updated through the event webpage: <a href="/en/about-chalmers/Chalmers-for-a-sustainable-future/sustainability-day2018/Pages/default.aspx">Chalmers Sustainability Day 2018</a><br /><br />For now, save the date!