Contextual Recommendations in Action - Bridging AI and Real-Life Economics

Description

Achieving our action requires to establish of a dialogue between experts in information sciences and consumer sciences. On the one hand, data scientists need to develop algorithms to reveal consumer preferences and produce recommendations. On the other hand, behavior economists need to design deployments controlled experiences according to principles established in cognitive psychology and behavioral economics. These efforts enable to develop theories on how people make choices when assisted by decision-making algorithms and to measure acceptability. These two parties have been working together to feed off each other and deploy large-scale testing through platform and plug-in development. We have been collecting data and metadata to conduct a robust measurement of the impact of algorithmic decisions on individuals and on society. To achieve that, we have designed and deployed large-scale experiments for the observation of consumption of behaviors in a controlled environment. These deployments are being conducted and tested according to new definitions of satisfaction and adoption measures, and of course the development of algorithms, tools and novel human-centric data exploration and analysis approaches.

Activities

This document summarizes our action and the resulting challenges. During the first year, we organized three meetings during which we designed our experiences on AdAnalyst, developed a methodology for deploying experiences and modeling preferences, and obtained the necessary authorizations from Data Protection Officers and analyzed the risks associated with GDPR.

The list of our publications is available here, as well as a lexicon that we have developed to allow us to work together between computer scientists, economists and lawyers.
We have also contributed to the Digital Platforms and Algorithmic Risks working group as part of the GdR Internet et Société: https://cis.cnrs.fr/plateformes-et-risques-algorithmiques/

Scientific publications

Le Monde Binaire


Ouvrages, Conferences, Journals

  • Faut-Il Avoir Peur Des Algorithmes ? Sihem Amer-Yahia, Juliette Sénéchal et Amélie Favreau. Ouvrage Humain et Numérique en Interaction, Editions CNRS 2020, à paraître
     
  • Facebook Ads Monitor: An Independent Auditing System for Political Ads on Facebook: Márcio Silva, Lucas Santos de Oliveira, Athanasios Andreou, Pedro Olmo Vaz de Melo, Oana Goga, Fabrício Benevenuto. WWW 2020 (The Web Conference). Honorable Mention Award, https://lig- membres.imag.fr/gogao/papers/pol_ads_WWW2020.pdf
     
  • Gain and Loss framing of incentives: encouraging individuals to provide a repetitive effort for small rewards: Penelope Buckley, Béatrice Roussillon and Sabrina Teyssier. ASFEE (Association Française d’économie expérimentale) June 2019 Toulouse
     
  • Enabling Decision Support Through Ranking and Summarization of Association Rules for TOTAL Customers: Idir Benouaret, Sihem Amer-Yahia, Senjuti Basu Roy, Christiane Kamdem Kengne, Jalil Chagraoui. Trans. Large Scale Data Knowl. Centered Syst. 44: 160-193 (2020)
     
  • Sihem Amer-Yahia, Shady Elbassuoni, Ahmad Ghizzawi, Ria Mae Borromeo, Emilie Hoareau, Philippe Mulhem: Fairness in Online Jobs: A Case Study on TaskRabbit and Google. EDBT 2020: 510-521
     
  • Fairness of Scoring in Online Job Marketplaces: Shady Elbassuoni, Sihem Amer-Yahia, Ahmad Ghizzawi. ACM Tansactions in Data Science, 2020 (to appear)
     
  • Exploring Fairness of Ranking in Online Job Marketplaces: Shady Elbassuoni, Sihem Amer-Yahia, Christine El Atie, Ahmad Ghizzawi, Bilel Oualha. EDBT 2019: 646-649
     
  • A Bi-Objective Approach for Product Recommendations: Idir Benouaret, Sihem Amer-Yahia, Christiane Kamdem Kengne, Jalil Chagraoui. BigData 2019: 2159-2168
     
  • An Efficient Greedy Algorithm for Sequence Recommendation: Idir Benouaret, Sihem Amer-Yahia, Senjuti Basu Roy. DEXA (1) 2019: 314-326
     
  • FaiRank: An Interactive System to Explore Fairness of Ranking in Online Job Marketplaces: Ahmad Ghizzawi, Julien Marinescu, Shady Elbassuoni, Sihem Amer-Yahia, Gilles Bisson. EDBT 2019: 582-585
     
  • Bulles de filtrage commercial: Tommaso Venturini, Paolo Frasca. Note pour une première expérience CNRS/Cdiscount
     
  • From filter to trending bubbles: Tommaso Venturini, Paolo Frasca
     
  • Plateformes numériques, algorithmes et société: explicabilité et effets sur PandHeMic Oct 31, 2019, https://pandhemic.hypotheses.org/1147

  • G. Fancello, A. Tsoukiàs, 'Learning urban capabilities from citizens' behaviour for urban planning’’, to appear in Socio-Economic Planning Sciences.
     
  • Richard A., Mayag B., Talbot F., Tsoukiàs A., Meinard Y., "What does it mean to provide decision support to a responsible and competent expert? The case of diagnostic decision support systems'', to appear in EURO Journal on Decision Processes.
     
  • G. Fancello, T. Conju, A. Tsoukiàs, "Mapping Walkability: a subjective value theory approach'', to appear in Socio-Economic Planning Sciences.
     
  • A. Colorni, A. Tsoukiàs, "Designing Alternatives for Decision Problems'', Journal of Multi-Criteria Decision Analysis, in press, 2020. DOI:10.1002/mcda.1709. Preliminary version.
     
  • I. Pluchinotta, R. Giordano, D. Zikos. T. Krüger, A. Tsoukiàs, "Integrating Problem Structuring Methods and Concept-Knowledge Theory for Advanced Policy Design: lessons from a case study in Cyprus'', in press in Journal of Comparative Policy Analysis, 2020. DOI: 10.1080/13876988.2020.1753512. Preliminary version.
     
  • O. Raboun, E. Chojnacki, C. Duffa, D. Rios-Insua, A. Tsoukiàs, "Spatial risk assessment in case of multiple nuclear release scenarios'', in press, Socio-Economic Planning Sciences, vol. 70, 2020. DOI: 10.1016/j.seps.2019.06.006. Preliminary version.
     
  • Giordano R., Pluchinotta I., Zikos D., Krueger T., Tsoukiàs A., "How to use Ambiguity in Problem Framing for Enabling Divergent Thinking: Integrating Problem Structuring Methods and Concept- Knowledge Theory'', in L. White, M. Kunc, K. Burger, J. Malpass (eds.), Behavioural Operational Research: a capabilities approach, Palgrave-MacMillan, 93 - 117, 2020. DOI: 10.1007/978-3-030-25405- 6_6. Preliminary version.