Research

You can use our plugin to insert parts from your activity report (raweb)service.

Presentation

Example : tyrex

Objectives

We work on the foundations of the next generation of data analytics and data-centric programming systems. These systems extend ideas from programming languages, artificial intelligence, data management systems, and theory. Data-intensive applications are increasingly more demanding in sophisticated algorithms to represent, store, query, process, analyse and interpret data. We build and study data-centric programming methods and systems at the core of artificial intelligence applications. Challenges include the robust and efficient processing of large amounts of structured, heterogeneous, and distributed data.

Last activity report : 2021

Results

New results

Algebraic Foundations for Distributed Query Evaluation

Big data programming frameworks are becoming increasingly important for the development of applications, for which performance and scalability are critical. In those complex frameworks, optimizing code by hand is hard and time-consuming, making automated optimization particularly necessary. In order to automate optimization, a prerequisite is to find suitable abstractions to represent programs; for instance, algebras based on monads or monoids to represent distributed data collections. Currently, however, such algebras do not represent recursive programs in a way which allows analyzing or rewriting them. In this paper, we extend a monoid algebra with a fixpoint operator for representing recursion as a first class citizen and show how it allows new optimizations. The fixpoint operator is suitable for modeling recursive computations with distributed data collections. We show that under reasonable conditions this fixpoint can be evaluated by parallel loops with one final merge rather than by a global loop requiring network overhead after each iteration. We also propose several rewrite rules, showing when and how filters can be pushed through recursive terms, and how to filter inside a fixpoint before a join. Experiments with the Spark platform illustrate performance gains brought by these systematic optimizations .

We have investigated the distributed evaluation of μ -RA queries. We present a system called Dist- μ -RA for the distributed evaluation of recursive graph queries. Dist- μ -RA builds on the recursive relational algebra and extends it with evaluation plans suited for the distributed setting. The goal is to offer expressivity for high-level queries while providing efficiency at scale and reducing communication costs. Experimental results on both real and synthetic graphs show the effectiveness of the proposed approach compared to existing systems .

Exploring Property Graphs with Recursive Path Patterns

We demonstrate a system for recursive query answering over property graphs. The novelty of the system resides in its ability to optimize and efficiently answer recursive path patterns in queries for property graphs. The system is based on a complete implementation of the μ -recursive relational algebra . It also includes parsers and compilers adapted for property graphs so that one can formulate, optimize and answer queries that navigate recursively along paths in property graphs. We demonstrate the system on three real datasets, including the exploration of chains of drug interactions .

Predictive Analytics for Healthcare

Federated learning becomes a prominent approach when different entities want to learn collaboratively a common model without sharing their training data. However, Federated learning has two main drawbacks. First, it is quite bandwidth inefficient as it involves a lot of message exchanges between the aggregating server and the participating entities. This bandwidth and corresponding processing costs could be prohibitive if the participating entities are, for example, mobile devices. Furthermore, although federated learning improves privacy by not sharing data, recent attacks have shown that it still leaks information about the training data. This paper presents a novel privacy-preserving federated learning scheme. The proposed scheme provides theoretical privacy guarantees, as it is based on Differential Privacy. Furthermore, it optimizes the model accuracy by constraining the model learning phase on few selected weights. Finally, as shown experimentally, it reduces the upstream and downstream bandwidth by up to 99.9% compared to standard federated learning, making it practical for mobile systems.

These results have been presented at the UAI 2021 conference .

Machine Learning, and in particular Federated Machine Learning, opens new perspectives in terms of medical research and patient care. Although Federated Machine Learning improves over centralized Machine Learning in terms of privacy, it does not provide provable privacy guarantees. Furthermore, Federated Machine Learning is quite expensive in term of bandwidth consumption as it requires participant nodes to regularly exchange large updates. This paper proposes a bandwidth-efficient privacy-preserving Federated Learning that provides theoretical privacy guarantees based on Differential Privacy. We experimentally evaluate our proposal for in-hospital mortality prediction using a real dataset, containing Electronic Health Records of about one million patients. Our results suggest that strong and provable patient-level privacy can be enforced at the expense of only a moderate loss of prediction accuracy.

These results have been presented at the CHIL 2021 conference .

Federated Learning allows distributed entities to train a common model collaboratively without sharing their own data. Although it prevents data collection and aggregation by exchanging only parameter updates, it remains vulnerable to various inference and reconstruction attacks where a malicious entity can learn private information about the participants’ training data from the captured gradients. Differential Privacy is used to obtain theoretically sound privacy guarantees against such inference attacks by noising the exchanged update vectors. However, the added noise is proportional to the model size which can be very large with modern neural networks. This can result in poor model quality. In this paper, compressive sensing is used to reduce the model size and hence increase model quality without sacrificing privacy. We show experimentally, using 2 datasets, that our privacy-preserving proposal can reduce the communication costs by up to 95% with only a negligible performance penalty compared to traditional non-private federated learning schemes. These results have been presented at the EuroS&P 2021 conference .

Early identification of patients at risk of developing complications during their hospital stay is currently a challenging issue in healthcare. Complications include hospital-acquired infections, admissions to intensive care units, and in-hospital mortality. Being able to accurately predict the patients’ outcomes is a crucial prerequisite for tailoring the care that certain patients receive, if it is believed that they will do poorly without additional intervention. We consider the problem of complication risk prediction, such as inpatient mortality, from the electronic health records of the patients. We study the question of making predictions on the first day at the hospital, and of making updated mortality predictions day after day during the patient’s stay. We develop distributed models that are scalable and interpretable. Key insights include analysing diagnoses known at admission and drugs served, which evolve during the hospital stay. We leverage a distributed architecture to learn interpretable models from training datasets of gigantic size. We test our analyses with more than one million of patients from hundreds of hospitals, and report on the lessons learned from these experiments.

Results presented at the 2018 International Conference on Data Science and Applications have been extended with a calibration study and measures for general and instance-level interpretations of the predictions .

Data Cleaning and Exchange

The automatic discovery of functional dependencies(FDs) has been widely studied as one of the hardest problems in data profiling. Existing approaches have focused on making the FD computation efficient while inspecting single relations at a time. In this paper, for the first time we address the problem of inferring FDs for multiple relations as they occur in integrated views by solely using the functional dependencies of the base relations of the view itself. To this purpose, we leverage logical inference and selective mining and show that we can discover most of the exact FDs from the base relations and avoid the full computation of the FDs for the integrated view itself, while at the same time preserving the lineage of FDs of base relations. We propose algorithms to speedup the inferred FD discovery process and mine FDs on-the-fly only from necessary data partitions. We present InFine(INferred FunctIoNal dEpendency), an end-to-end solution to discover inferred FDs on integrated views by leveraging provenance information of base relations. Our experiments on a range of real-world and synthetic datasets demonstrate the benefits of our method over existing FD discovery methods that need to rerun the discovery process on the view from scratch and cannot exploit lineage information on the FDs. We show that InFine outperforms traditional methods necessitating the full integrated view computation by one to two order of magnitude in terms of runtime. It is also the most memory efficient method while preserving FD provenance information using mainly inference from base table with negligible execution time.

These results will be presented at the ICDE 2022 conference .

We study the explainability of automated data cleaning pipelines and propose CLeanEX, a solution that can generate explanations for the pipelines automatically selected by an automated cleaning system, given it can provide its corresponding cleaning pipeline search space. We propose meaningful explanatory features that are used to describe the pipelines and generate predicate-based explanation rules. We compute quality indicators for these explanations and propose a multi-objective optimization algorithm to select the optimal set of explanations for user-defined objectives. Preliminary experiments show the need for multi-objective optimization for the generation of high-quality explanations that can be either intrinsic to the single selected cleaning pipeline or relative to the other pipelines that were not selected by the automated cleaning system. We also show that CLeanEX is a promising step towards generating automatically insightful explanations, while catering to the needs of the user alike. .

Exchanging data between data sources is a fundamental problem in many data science and data integration tasks. In this paper, we focus on the data exchange problem in the presence of privacy constraints on the source data, which has been disregarded in the literature to date. By leveraging a logical privacy-preservation paradigm, the privacy restrictions are expressed as a set of policy views representing the information that is safe to expose over all instances of the source in order to exchange them with the target. We introduce a protocol that provides formal privacy guarantees and is data-independent, i.e., under certain criteria, it guarantees that the mappings leak no sensitive information independently of the instances lying in the source. Moreover, we design an algorithm for repairing an input mapping w.r.t. a set of policy views, in cases where the input mapping leaks sensitive information. We show that the repairing can build upon hard-coded and learning-based user preference functions and we show the trade-offs. Our empirical evaluation shows that repairing mappings is quite efficient, leading to repairing sets of 300 s-t tgds in an average time of 5s on a commodity machine. It also shows that the repairing based on learning is robust and has comparable runtimes with the hard-coded one .

You can write want you want/need on this page by using HTML tags in the text editor or use the visual editor.

  • Research direction 1

    …….

  • Research direction 2

    ……….

  • Research direction 3

    ……….

  • Comments are closed.