Design of Experimental Search & Information REtrieval Systems Omar Alonso Gianmaria Silvello Microsoft University of Padua Mountain View, CA, USA Padua, Italy omalonso@microsoft.com silvello@dei.unipd.it SCOPE OF THE CONFERENCE FORWARD TO THE PROCEEDINGS DESIRES (Design of Experimental Search & Information REtrieval These proceedings contain the full papers and short papers selected Systems) is a systems-oriented conference, complementary in its for presentation at the First Biennial Conference on Design of Exper- mission to the mainstream Information Access and Retrieval confer- imental Search & Information REtrieval Systems (DESIRES 2018). 1 ences like SIGIR, ECIR, and other symposiums focusing on specific The event is organized under the Bertinoro international Center aspects of IR such as ICTIR or CHIIR, emphasizing the innovative for informatics (BiCi) 2 umbrella. BiCi is an association whose mis- technological aspects of search and retrieval systems. sion is to foster cutting-edge research and advanced education in DESIRES gathers researchers and practitioners from both academia Computer Science. The conference was held from August 28 to and industry to discuss the latest innovative and visionary ideas August 31, 2018, at the University Residential Centre of Bertinoro in the field. The goal of this new conference is to provide the IR (Ce.U.B.) 3 in Bertinoro, Italy. community with a venue for presenting innovative search systems DESIRES looked for original work that could not be submitted architectures, as well as a publication opportunity. DESIRES does in parallel to another venue. DESIRES is a single-track conference not compete with the established conferences presenting rigorous focused on discussion. In order to encourage authors to submit only treatises in established areas; rather its goal is to air radically new their best work, each person could be an author or co-author of ideas. only a single full paper. DESIRES mainly encourages papers about innovative and risky Full papers presented at DESIRES may lack rigorous frameworks, information access and retrieval system ideas, systems-building ex- simulations of performance, or traditional evaluations but present perience and insight, resourceful experimental studies, provocative a radical departure from conventional approaches that enables new position statements, new application domains. DESIRES also wel- applications. comes contributions focusing on implementation details, successful Abstracts present ideas that are still in early stages for a full or failed reproducibility attempts, technological breakthroughs, and paper or challenge the current trends in IR. Any author of a full new uses of old ideas. paper could additionally submit one abstract. Abstracts have a Traditionally, program committees in our IR field reward scholar- single author and were presented in a gong-show style. DESIRES ship on narrow ideas, operate by consensus, discard “loosey-goosey” received a total of 33 submissions in two categories: 20 full papers papers including half-baked ideas and submissions that resemble and 13 abstracts. war stories from the field. The major IR conferences usually reject All full papers were reviewed by at least three members of an such submissions because they are not scientific. However, these international Program Committee formed by experts from industry are often the very papers that offer long-term value to the field and academia. Of the full papers submitted to the conference, 13 and should be widely disseminated. DESIRES values innovation, were accepted for oral presentation. All the the abstracts were experience-based insight, and vision. reviewed by at least two reviewers and they were all accepted to be DESIRES is a single-track conference and there are three kinds presented to the conference. The accepted contributions represent of contributions: the state of the art in information retrieval, cover a diverse range of topics, propose new uses for IR techniques, querying, personal Research papers Papers usually lack rigorous frameworks, search, expert search, visual systems, IR research tools, teaching in simulations of performance, or prototype implementations IR, evaluation, NLP and collaborative search. but present a radical departure from conventional approaches There were 49 unique authors with papers or abstracts accepted that enables new applications. at DESIRES 2018 with the following geographical distribution (affil- Prototypes The prototype descriptions generally are a de- iation): 12 authors from the USA (24.5%), 9 authors from Germany tailed report on successes and mistakes. (18%), 7 authors from the UK (14%), 7 authors from The Netherlands Abstracts Ideas that are too half-baked for a paper or demo (14%), 5 authors from Australia (10%), 4 authors from Italy (8%), 2 proposal are good candidates for an abstract. authors from Qatar (4%) and one author from France, Sweden and DESIRES will run every other year in a retreat-like fashion (around Switzerland (2%). the time of the European Summer School in Information Retrieval There are three keynotes: (ESSIR)); ESSIR runs on odd years, DESIRES on even years. 1 http://desires.dei.unipd.it/ DESIRES 2018, August 2018, Bertinoro, Italy 2 http://www.bici.eu/index.html © 2018 Copyright held by the author(s). 3 http://www.ceub.it/?lang=en DESIRES 2018, August 2018, Bertinoro, Italy Omar Alonso and Gianmaria Silvello • Computing without Servers, V8, Rocket Ships, and Other Batsh*t Claudia Hauff, Delft University of Technology, The Netherlands Crazy Ideas in Data Systems, Jimmy Lin, University of Wa- Nicola Ferro, University of Padua, Italy terloo, Canada. Evgenyi Gabrilovich, Google, USA • Learning to Rank at Bloomberg - From Theory to Production, Krishna Gummadi, MPI-SWS, Germany Diego Ceccarelli, Bloomberg, UK. Vasileios Kandylas, Microsoft, USA • Balancing Efficiency and Effectiveness Trade-Offs in Large Bhaskar Mitra, Microsoft, UK Scale Multi-Stage Search Engines, Shane Culpepper, RMIT Nicola Montecchio, Spotify, Germany University, Australia. Edgar Meij, Bloomberg, UK There is one industry panel. The goal of the panel is to discuss Jussi Karlgren, Gavagai & KTH, Sweden topics that are of importance in industrial settings and may not be Emre Kiciman, Microsoft, USA of significance in academia. Moreover, another goal is to examine Andreas Rauber, Vienna University of Technology, Austria problems, emergent trends, and their implications for the future of Tony Russell-Rose, 2dSearch, UK the field. Tetsuya Sakai, Waseda University, Japan The speakers of the panel are: Mark Sanderson, RMIT, Australia Sunita Sarawagi, IIT Bombay, India • Emre Kiciman, Microsoft Ian Soboroff, NIST, USA • David Lewis, Cyxtera Paul Thomas, Microsoft, USA • Mark Najork, Google Andrew Trotman, eBay • Tony Russel Rose, UXLabs Arjen de Vries, Radboud University, The Netherlands DESIRES 2018 has been held under the patronage of the Depart- Ingmar Weber, QCRI, Qatar ment of Information Engineering of the University of Padua. Wouter Weerkamp, 904 Labs, The Netherlands Finally, DESIRES 2018 would have not been possible without Cong Yu, Google Research, USA the financial support from Bloomberg (silver sponsor) and Google Justin Zobel, University of Melbourne, Australia (bronze sponsor). Local Organization Chair ORGANIZATION Roberta Partisani, Ce.U.B., Italy DESIRES has been conceived by the general chairs with the support and suggestions of the advisory board. Advisory board members have been selected amongst the top senior members of the IR community seeking for balance between industry and academia. We thanks the advisory board members for their work and contribution to the conference. We thank all Program Committee members for their time and effort in ensuring the high quality of the DESIRES 2018 program. General Chairs Omar Alonso, Microsoft, USA Gianmaria Silvello, University of Padua, Italy Advisory Board Maristella Agosti, University of Padua, Italy Ricardo Baeza-Yates, NTENT, USA; UPF, Spain; Univ. de Chile Bruce Croft, University of Massachusetts, USA Susan Dumais, Microsoft, USA Norbert Fuhr, University of Duisburg-Essen, Germany Donna Harman, NIST, USA Kalervo Järvelin, University of Tampere, Finland Marc Najork, Google, USA Jan Pedersen, eBay, USA Gerhard Weikum, MPI, Germany Program Committee James Allan, University of Massachusetts, USA Barla Cambazoglu, NTENT, USA Diego Ceccarelli, Bloomberg, UK Jeffrey Dalton, University of Glasgow, UK