• Artificial Intelligence in the Industry

    Criteo

    Join us at Criteo HQ Paris to gain insight on the current challenges faced by the AI-based tech industry. We will be hosting worldwide top-notch speakers working to build the online applications of tomorrow. ⌚ Agenda : 6:30 pm : welcome 7:00 pm : talks 9:00 pm : networking _______________________________________________________________________________________ 🎤 Ricardo Baeza-Yates, CTO at Ntent 📃 Semantic Search and Query Intent Prediction Semantic search lies in the cross roads of information retrieval and natural language processing and is the current frontier of search technology. A main component is predicting the intention behind the query, which implies doing semantic query understanding. This process implies the same semantic processing as a document. After, based on all this information, we have to predict one or more possible intentions with a certain probability, which is particularly important for ambiguous queries. These scores will be one of the inputs for the final semantic ranking, a key second component. Semantic ranking refers to ranking search results using semantic information. In a standard search engine, a rank is computed by using signals or features coming from the search query, from the documents in the collection being searched and from the search context, such as the language and device being used. In our case we add semantic relations between the entities and concepts found in the query was the same objects in the documents, that will come from different data sources. For this we use machine learning in several stages. The first stage selects the data sources that we should use to answer the query. In the second stage, each data source generates a set of answers using ``learning to rank.'' The third and final stage ranks these data sources, selecting and ordering the intentions as well as the answers inside each intention (e.g., news) that will appear in the final composite answer. All these stages are language independent, but may use language dependent features. 🎤 Mounia Lalmas, Director of Research at Spotify 📃 Personalizing the listening experience When users interact with a recommendation system, they leave behind fine grained traces of interaction signals, which contain valuable information that could help gauging user satisfaction. Quantifying such a notion of satisfaction from implicit signals involves 1) understanding the diverse needs of users, 2) their expectations of what is a successful streaming session, 3) the recommender system goal to provide a personalized experience that also allows for discovery. This talk will describe methods explored to provide a more informed definition of satisfaction. 🎤 Patrick Gallinari, Distinguished Researcher at Criteo 📃 Learning visual context representations Characterization of objects in images most often relies on the intrinsic properties of the objects, i.e. their visual appearance. The context in which these objects do appear is an important source of information but is usually ignored. We present two investigations demonstrating the importance of visual context. One concerns learning multimodal words representation using both textual information and the visual context of corresponding objects. The second one addresses the problem of zero shot learning for object recognition in images. 🎤 Diego Saez Trumper, Research Scientist at Wikimedia 📃 Wikimedia Public (Research) Resources The Wikimedia Foundation's mission is to disseminate open knowledge effectively and globally. In keeping with this mission, the Wikimedia Foundation support research in areas that benefit the Wikimedia community. We aim to make any work with our support openly available to the public. At the same time that we do a minimalist user data collection, all the material (text and multimedia) available in our projects is public and reusable by everybody.