Research#
My research started with a Ph.D. in Computer Science at the University of Trento (2009–2014), advised by Prof. Themis Palpanas, on modeling and querying data with uncertainty: time series and streams whose values are noisy or incomplete, graphs whose edges may or may not exist, and the data cleaning that comes before any of it. Dissertation: Modeling and Querying Data Series and Data Streams with Uncertainty. Along the way I spent six months as a visiting researcher with the data management and analytics groups at the IBM T.J. Watson Research Center (USA) and three months at the Qatar Computing Research Institute (Qatar). The work appeared in SIGMOD, VLDB, EDBT, SSDBM, KAIS and DKE.
Consulting keeps raising questions worth writing up, on verifiable AI inference, forecasting with calibrated uncertainty and evaluation, and I publish those as preprints and reports. All publications are listed below, newest first, each linking to the publisher's version or the official PDF.
Google Scholar profile Consulting experience
Themes#
-
Uncertain time series
How to compare and search data series when each value carries uncertainty: similarity measures, correlation-aware distances and top-k nearest neighbor search. Five papers, 2011–2017.
-
Uncertain graphs
Classifying nodes in networks whose edges exist only with a probability, with Charu Aggarwal at IBM Research. Two papers, 2014 and 2018.
-
Data streams
Sliding windows over uncertain streams in IBM's System S, and finding frequent items in ad-hoc recent time windows. Two papers, 2013 and 2015.
-
Data cleaning and verifiable inference
NADEEF, a commodity data cleaning system built at the Qatar Computing Research Institute with Ihab Ilyas (SIGMOD 2013), and SPEX, a statistical proof of execution for AI inference (2025 preprint).
Publications#
| Year | Title | Authors | Venue |
|---|---|---|---|
| 2025 | Statistical Proof of Execution (SPEX) | Michele Dallachiesa, Antonio Pitasi, David Pinger, Josh Goodbody, Luis Vaello | arXiv preprint |
| 2018 | Improving Classification Quality in Uncertain Graphs | Michele Dallachiesa, Charu Aggarwal, Themis Palpanas | ACM Journal of Data and Information Quality (JDIQ) |
| 2017 | Similarity Using Correlation-Aware Measures | Katsiaryna Mirylenka, Michele Dallachiesa, Themis Palpanas | International Conference on Scientific and Statistical Database Management (SSDBM), Chicago, USA |
| 2017 | Correlation-Aware Distance Measures for Data Series | Katsiaryna Mirylenka, Michele Dallachiesa, Themis Palpanas | International Conference on Extending Database Technology (EDBT), Italy |
| 2015 | Top-k Nearest Neighbor Search in Uncertain Data Series | Michele Dallachiesa, Themis Palpanas, Ihab F. Ilyas | Proceedings of the VLDB Endowment (PVLDB) 8(1) |
| 2015 | Sliding Windows over Uncertain Data Streams | Michele Dallachiesa, Gabriela Jacques-Silva, Bugra Gedik, Kun-Lung Wu, Themis Palpanas | Knowledge and Information Systems (KAIS) 45 |
| 2014 | Node Classification in Uncertain Graphs | Michele Dallachiesa, Charu Aggarwal, Themis Palpanas | International Conference on Scientific and Statistical Database Management (SSDBM), Aalborg, Denmark |
| 2013 | NADEEF: A Commodity Data Cleaning System | Michele Dallachiesa, Amr Ebaid, Ahmed Eldawy, Ahmed Elmagarmid, Ihab Ilyas, Mourad Ouzzani, Nan Tang | ACM SIGMOD International Conference on Management of Data (SIGMOD), New York, USA |
| 2013 | Identifying Streaming Frequent Items in Ad-hoc Recent Time Windows | Michele Dallachiesa, Themis Palpanas | Data & Knowledge Engineering (DKE) |
| 2012 | Uncertain Time-Series Similarity: Return to the Basics | Michele Dallachiesa, Besmira Nushi, Katsiaryna Mirylenka, Themis Palpanas | Proceedings of the VLDB Endowment (PVLDB), Istanbul, Turkey |
| 2011 | Similarity Matching for Uncertain Time Series: Analytical and Experimental Comparison | Michele Dallachiesa, Besmira Nushi, Katsiaryna Mirylenka, Themis Palpanas | ACM SIGSPATIAL Workshop on Querying and Mining Uncertain Spatio-Temporal Data (QUeST), Chicago, USA |
Prof. Themis Palpanas, Ph.D. advisor, Université Paris Cité
"Michele is able to grasp new ideas and concepts, and in general, learn fast. The leading abilities of Michele are evident in his work: he is capable of independent thinking, and of delivering novel and effective solutions to hard problems. He combines a very solid theoretical background with excellent practical skills."
From research to practice#
The same questions run through my consulting work: how much to trust a forecast, how to put calibrated bounds on a prediction, and how to evaluate a system before relying on it. Conformal prediction in railway cost forecasting, explainable risk rankings for megaprojects and urban forests, leave-one-out testing of infrastructure cost models, and the evaluation framework for voice agents all start from there.
Have a problem with this shape?
If your decision depends on a forecast, a ranking or a model whose trustworthiness you cannot yet measure, that is the kind of work I take on.