Giovani in un'ora - Ciclo di seminari - Quarta parte
- Day - Time: 30 November -0001, h.00:00
- Place: Area della Ricerca CNR di Pisa - Room: C-29
Abstract: To be profitable, search engines need to be both effective and efficient when processing queries. A search engine is effective if its results (lists of documents) contain documents relevant to the queries in their first positions. A search engine is efficient if it generates the query results in a little amount of time. There are several approaches to improve effectiveness or efficiency. For instance, a search engine can rewrite users' queries to enhance effectiveness, e.g., by applying proximity models or by expanding the query with additional keywords. However, rewritings that benefit effectiveness often have a negative impact on efficiency. Differently, efficiency can be improved by early terminating the query processing, i.e., by processing only a subset of the indexed documents which match the query. Reinforcement learning can be used by the search engine to learn when early termination leads to no or just small degradation in effectiveness. In this work, we want to investigate whether reinforcement learning can be successfully applied to learn how to efficiently and effectively rewrite queries.
Davide Basile - "An Experience in Applying Formal Methods to Railway Systems"
Abstract: CENELEC EN 50128 standard for the development of software for railway control and protection systems specifically mentions formal methods as highly recommended practices for software systems to be certified at Safety Integrity Levels (SIL) 3 and 4. In this seminar I will briefly describe recent efforts in applying formal methods techniques to railway systems in the context of two projects in which I have been involved in the last year: Tuscany Region project SISTER and ASTRail project, which received funding from the Shift2Rail Joint Undertaking under the European Union's Horizon 2020 research and innovation programme under Grant Agreement No. 777561.
Riccardo Guidotti - "Clustering individual transactional data for masses of users"
Abstract: Mining a large number of datasets recording human activities for making sense of individual data is the key enabler of a new wave of personalized knowledge-based services. We analyzed the problem of clustering individual transactional data for a large mass of users. Transactional data is a very pervasive kind of information that is collected by several services, often involving huge pools of users. We propose txmeans, a parameter-free clustering algorithm able to efficiently partitioning transactional data in a completely automatic way. Txmeans is designed for the case where clustering must be applied on a massive number of different datasets. A deep experimentation on both real and synthetic datasets shows the practical effectiveness of txmeans for the mass clustering of different personal datasets. Finally, we present a personal cart assistant application based on txmeans.