|Nick Craswell, Neural Models for Full Text Search: Could the Improvements Add Up?|
|10:55||B2.15||Session A [Session Chair: Leif Azzopardi]|
|10:55 Shane Culpepper, Charles Clarke and Jimmy Lin, Dynamic Cutoff Prediction in Multi-Stage Retrieval Systems|
|11:20 Jimmy, Guido Zuccon and Bevan Koopman, Boosting Titles does not Generally Improve Retrieval Effectiveness|
|11:45 Sheng Wang, Zhifeng Bao, Shane Culpepper, Timos Sellis, Mark Sanderson and Munkh-Erdene Yadamjav, Interactive Trip Planning Using Activity Trajectories|
|13:15||B2.14||Keynote Session (Shared with ALTA)|
|14:05||B2.15||Session B [Session Chair: Mark Sanderson]|
|14:05 Tadele Damessie, Falk Scholer and Shane Culpepper, The Influence of Topic Difficulty, Relevance Level, and Document Ordering on Relevance Judging|
|14:30 Andrew Trotman and Jimmy Lin, In Vacuo and In Situ Evaluation of SIMD Codecs|
|14:55 Alistair Moffat, Judgment Pool Effects Caused by Query Variations|
|15:10 Laurence Park, Uncertainty in Rank-Biased Precision|
|16:00||B2.15||Session C [Session Chair: Alexandra Uitdenbogerd]|
|16:00 Benoit Potvin, Roger Villemaire and Ngoc-Tan Le, A Position-Based Method for the Extraction of Financial Information in PDF Documents|
|16:25 Sunghwan Mac Kim, Stephen Wan and Cecile Paris, Occupational Representativeness in Twitter|
|18:00||Conference Drinks and Dinner at Caulfield Glasshouse|
|9:10||B2.15||Session D [Session Chair: Paul Thomas]|
|9:10 Dinesha Chathurani Nanayakkara Wasam Uluwitige, Shlomo Geva, Guido Zuccon, Vinod Chandran and Timothy Chappell, Effective User Relevance Feedback for Image Retrieval with Image Signatures|
|9:35 Wern Han Lim, Mark Carman and Sze-Meng Wong, Estimating Domain-Specific User Expertise for Answer Retrieval in Community Question-Answering Platforms|
|10:00 Gaya K. Jayasinghe, Sarvnaz Karimi and Melanie Ayre, Evaluation of Retrieval Algorithms for Expertise Search|
|10:40||B2.15||Poster Boosters Session [Session Chair: Laurianne Sitbon]|
|Jaewon Kim, Paul Thomas, Ramesh Sankaranarayana, Tom Gedeon and Hwan-Jin Yoon. Understanding Mobile Web Search Behaviour|
|Ameer Albahem. Dynamic Diversification in Information Retrieval|
|Riyad Alrihieli. Privacy Settings on Social Networks|
|Shiwei Zhang and Xiuzhen Zhang. Factors Affecting the Informativeness of Disaster Tweets|
|Laurianne Sitbon. Designing new information access technologies with People with Intellectual Disability|
|Xiaolu Lu, Alistair Moffat and Shane Culpepper. Estimating Deep Metric Scores Using Shallow Judgment Pools|
|Manuel Cebrian, Time-Critical Social Mobilization|
|14:30||B2.15||Session E [Session Chair: Laurence Park]|
|14:30 Evi Yulianti, Ruey-Cheng Chen, Falk Scholer and Mark Sanderson, Using Semantic and Context Features for Answer Summary Extraction|
|14:45 Vincent Au, Paul Thomas and Gaya K. Jayasinghe, Query-Biased Summaries for Tabular Data|
|Michael Cameron, Analyzing user data at travel start-up Rome2rio|
|16:00||B2.15||Best paper announcement|
|16:30||B2.15||ADCS 2016 Closing|
Note that full paper presentations are allocated 25 minutes (including question time) and short paper presentations are allocated 15 minutes (including question time). Authors of abstracts are allocated 5 minutes each in the poster booster session.
Information on tutorials on 5th of December can be found on ALTA 2016 Tutorial page.
Title:Neural Models for Full Text Search: Could the Improvements Add Up?
Abstract: Neural word embeddings and deep neural networks may yet give us worthwhile performance gains on standard IR tasks such as TREC ad hoc, and become a standard part of the IR toolkit. On the other hand, improvements may not be large enough or reliable enough, and the approach may fail. As a research community, our progress in the area may be limited by a lack of large-scale training data. We also need to apply our most rigorous tests, such as blind testing at TREC. This talk will cover some recent progress in neural models for full text search, describe the training data requirements, and discuss what it would take to really prove that the neural models are worthwhile.
Title: Time-Critical Social Mobilization
Abstract:This talk explores the physical and behavioral limits of crowds-assembly for problem solving, by following a number of real-world experiments where we utilized social media to mobilize the masses in tasks of unprecedented complexity. From finding red weather balloons to locating people in distant cities, to reconstructing shredded documents, the power of crowdsourcing is real, but so are exploitation, sabotage, and hidden biases that undermine the power of crowds.