Home » Node » 27626

Seminario pubblico Giuseppe Antonio Di Luna

speaker DIAG: 
English translation unavailable for diluna@diag.uniroma1.it.
Data dell'evento: 
Tuesday, 30 January, 2024 - 12:00
Luogo: 
Aula A5
Contatto: 
lazzeretti@diag.uniroma1.it

In ottemperanza ai requisiti previsti dalla procedura valutativa ai fini del passaggio a professore di II Fascia SC 09/H1 SSD ING-INF/05 - Dipartimento di Ingegneria Informatica Automatica e Gestionale "A. Ruberti",

martedì 30 gennaio 2024 alle ore 12:00, in aula A5 si terrà il seminario di Giuseppe Antonio Di Luna che illustrerà le sue attività di ricerca svolte e in corso di svolgimento.

Il seminario sarà anche trasmesso in modalità telematica su Zoom. Per partecipare da remoto connettersi all'indirizzo seguente:

https://uniroma1.zoom.us/j/82300723224

 

Titolo: Theory of Dynamic Networks and Practice of Binary Analysis 

 

Abstract: The talk will provide an overview of Di Luna's recent research activities, ranging from theoretical work in distributed computing to practical applications in compiler testing and cybersecurity.

 

The first part of the talk will concentrate on a series of recent studies concerning the computability of functions in anonymous dynamic networks. In dynamic networks, unlike static networks, the topology connecting processes changes continuously and unpredictably. This paradigm models several real-world scenarios, such as a peer-to-peer network of smartphones. A seminal paper presented at STOC 2010 demonstrated that when processes have unique identifiers, it is possible to compute any function of the process inputs in time linear to the network size. However, the situation is different for anonymous processes without unique identifiers. This limitation is common in scenarios where privacy is a concern, such as COVID-19 tracking apps. In such cases, all known deterministic algorithms for basic, non-trivial computations, such as counting the number of processes or calculating the average of inputs, have at least quartic complexity in terms of network size. This complexity made deterministic computation seem impractical in real-world scenarios. Through a series of works conducted during my RTD-B at DIAG, we introduced a new combinatorial data structure, the history tree (FOCS 2022). This structure fully captures the dynamics of anonymous networks, enabling us to develop definitive linear-time algorithms for computing all computable functions in anonymous dynamic networks (FOCS 2022, DISC 2023). These algorithms have enhanced the state of the art in several areas, including the well-known average consensus problem, showing that deterministic computation in anonymous dynamic networks can be done efficiently. 

 

In the second part of the talk, Dr. Di Luna will discuss his research in practical systems, focusing on the correctness and completeness of debug information in optimized binaries (ASPLOS 2021, ASPLOS 2023) and on techniques using deep neural networks for automated binary analysis.
 

 

Short Bio: Giuseppe Antonio Di Luna got his Ph.D. from Sapienza University of Rome in 2015. After his Ph.D. he did a postdoc at the University of Ottawa, working on fault-tolerant distributed algorithms, distributed robotics, and algorithm design for programmable particles. In 2018 he started a postdoc at the Aix-Marseille University, where he worked on dynamic graphs. He has been a postdoctoral fellow at Sapienza funded by the AXA fellowship and performing research on applying NLP techniques to the binary analysis domain. Currently, is an RTD-B at the DIAG department working on theory of distributed computing and practical aspects of systems. 

gruppo di ricerca: 
© Università degli Studi di Roma "La Sapienza" - Piazzale Aldo Moro 5, 00185 Roma