Projects

In this project, we work on providing a privacy-preserving biometric solution for humanitarian aid distribution. The project seeks to understand the requirements of aid distribution in emergency situation and design a solution that enables the use of biometrics without endangering the beneficiaries that need access to aid.

State-of-the-art architectures for modulation recognition are typically based on deep learning models. However, recently these models have been shown to be quite vulnerable to very small and carefully crafted perturbations, which pose serious questions in terms of safety, security, or performance guarantees at large. While adversarial training can improve the robustness of the network, there is still a large gap between the performance of the model against clean and perturbed samples. Based on recent experiments, the data used during training could be an important factor in the susceptibility of the models. Thus, the objective of this project is to research the effects of proper data selection, cleaning and preprocessing of the samples used during training on robustness.

After 18 months of responding to the COVID-19 pandemic, there is still no agreement on the optimal combination of mitigation strategies. The efficacy and collateral damage of pandemic policies are dependent on constantly evolving viral epidemiology as well as the volatile distribution of socioeconomic and cultural factors. This study proposes a data-driven approach to quantify the efficacy of the type, duration, and stringency of COVID-19 mitigation policies in terms of transmission control and economic loss, personalised to individual countries.

The objective of the TMM project is to identify, at an early stage, the risks associated with new technologies and develop solutions to ward off such threats. It also aims to assess existing products and applications to pinpoint vulnerabilities. In that process, artificial intelligence and machine learning will play an important part. The main goal of this project is to automatically identify technology offerings of Swiss companies especially in the cyber security domain. This also includes identifying key stakeholders in these companies, possible patents, published scientific papers.

Recently, deep neural networks have been applied in many different domains due to their significant performance. However, it has been shown that these models are highly vulnerable to adversarial examples. Adversarial examples are slightly different from the original input but can mislead the target model to generate wrong outputs. Various methods have been proposed to craft these examples in image data. However, these methods are not readily applicable to Natural Language Processing (NLP). In this project, we aim to propose methods to generate adversarial examples for NLP models such as neural machine translation models in different languages. Moreover, through adversarial attacks, we mean to analyze the vulnerability and interpretability of these models.

The objective of the TMM project is to identify, at an early stage, the risks associated with new technologies and develop solutions to ward off such threats. It also aims to assess existing products and applications to pinpoint vulnerabilities. In that process, artificial intelligence and machine learning will play an important part. The main goal of this project is to automatically identify technology offerings of Swiss companies especially in the cyber security domain. This also includes identifying key stakeholders in these companies, possible patents, published scientific papers.

Modulation recognition state-of-the-art architectures use deep learning models. These models are vulnerable to adversarial perturbations, which are imperceptible additive noise crafted to induce misclassification, posing serious questions in terms of safety, security, or performance guarantees at large. One of the best ways to make the model robust is to use adversarial learning, in which the model is fine-tuned with these adversarial perturbations. However, this method has several drawbacks. It is computationally costly, has convergence instabilities and it does not protect against multiple types of corruptions at the same time. The objective of this project is to develop improved and effective adversarial training solutions that tackle these drawbacks.

We consider a web-scale application within a datacenter that comprises of hundreds of software components, deployed on thousands of servers. These versatile components communicate with each other via Remote Procedure Calls (RPCs) with the cost of an individual RPC service typically measured in microseconds. The end-user performance, availability and overall efficiency of the entire system are largely dependent on the efficient delivery and scheduling of these RPCs. We propose to make RPC first-class citizens of datacenter deployment. This requires a revisitation of the overall architecture, application API, and network protocols. We are also building the tools that are necessary to scientifically evaluate microsesecond-scale services.

The overall goal of this project is to develop methods for monitoring, modeling, and modifying dietary habits and nutrition based on large-scale digital traces. We will leverage data from both EPFL and Microsoft, to shed light on dietary habits from different angles and at different scales.
Our agenda broadly decomposes into three sets of research questions: (1) Monitoring and modeling, (2) Quantifying and correcting biases and (3) Modifying dietary habits.
Applications of our work will include new methods for conducting population nutrition monitoring, recommending better-personalized eating practices, optimizing food offerings, and minimizing food waste.

P4 (Predictive, Preventive, Personalized and Participatory) medicine is called to revolutionize healthcare by providing better diagnoses and targeted preventive and therapeutic measures. In order to enable effective P4 medicine, DPPH defines an optimal balance between usability, scalability and data protection, and develops required computing tools. The target result of the project will be a platform composed of software packages that seamlessly enable clinical and genomic data sharing and exploitation across a federation of medical institutions across Switzerland. The platform is scalable, secure, responsible and privacy-conscious. It can seamlessly integrate widespread cohort exploration tools (e.g., i2b2 and TranSMART).

We develop a new method that detects jumps nonparametrically in financial time series and significantly outperforms the current benchmark on simulated data. We use a long short- term memory (LSTM) neural network that is trained on labelled data generated by a process that experiences both jumps and volatility bursts. As a result, the network learns how to disentangle the two. Then it is applied to out-of-sample simulated data and delivers results that considerably differ from the benchmark: we obtain fewer spurious detection and identify a larger number of true jumps. When applied to real data, our approach for jump screening allows to extract a more precise signal about future volatility.

We develop a methodology for detecting asset bubbles using a neural network. We rely on the theory of local martingales in continuous-time and use a deep network to estimate the diffusion coefficient of the price process more accurately than the current estimator, obtaining an improved detection of bubbles. We show the outperformance of our algorithm over the existing statistical method in a laboratory created with simulated data. We then apply the network classification to real data and build a zero net exposure trading strategy that exploits the risky arbitrage emanating from the presence of bubbles in the US equity market from 2006 to 2008. The profitability of the strategy provides an estimation of the economical magnitude of bubbles as well as support for the theoretical assumptions relied on.

I propose a new tool to characterize the resolution of uncertainty around FOMC press conferences. It relies on the construction of a measure capturing the level of discussion complexity between the Fed Chair and reporters during the Q&A sessions. I show that complex discussions are associated with higher equity returns and a drop in realized volatility. The method creates an attention score by quantifying how much the Chair needs to rely on reading internal documents to be able to answer a question. This is accomplished by building a novel dataset of video images of the press conferences and leveraging recent deep learning algorithms from computer vision. This alternative data provides new information on nonverbal communication that cannot be extracted from the widely analyzed FOMC transcripts. This paper can be seen as a proof of concept that certain videos contain valuable information for the study of financial markets.

Armed conflicts, violence and migration are causing large scale separation of family members, dislocation of family links and missing persons. People must receive help to know what happened to reconnect to their loved ones as rapidly as possible. The ICRC and LSIR through its partnership have set themselves a challenge to analyse publicly available data through analytics techniques to identify missing persons that would arguably not have been identified using current, conventional methods. The goal of this project is to facilitate the search for missing individuals by building scalable, accurate systems tailored for that purpose.

Aircraft and their ground counterparts have been communicating via the ACARS data-link protocol for more than five decades. Researchers discovered that some actors encrypt ACARS messages using an insecure, easily reversible encryption method. In this project, we propose BRUTUS, a decision-support system that support human analysts to detect the use of insecure ciphers in the ACARS network in an efficient and scalable manner. We propose and evaluate three different methods to automatically label ACARS messages that are likely to be encrypted with insecure ciphers.

Cyber security information is often extremely sensitive and confidential, it introduces a tradeoff between the benefits of improved threat-response capabilities and the drawbacks of disclosing national-security-related information to foreign agencies or institutions. This results in the retention of valuable information (a.k.a. as the free-rider problem), which considerably limits the efficacy of data sharing. The purpose of this project is to resolve the cybersecurity information-sharing tradeoff by enabling more accurate insights on larger amounts of more relevant collective threat-intelligence data.
This project will have the benefit of enabling institutions to build better models by securely collaborating with valuable sensitive data that is not normally shared. This will expand the range of available intelligence, thus leading to new and better threat analyses and predictions.

In this collaboration (structured in two projects) we develop an automated tool to flag messages sent by planes which are suspicious of using weak encryption mechanisms. We mainly focus on detecting the use of classical ciphers like substitution and transposition ciphers. The tool flags messages and identifies the family of ciphers. We also aim to develop automated decryption techniques for the weakest ciphers.

In this report we consider several real-life scenarios that may provoke causal research questions. As we introduce concepts in causal inference, we reference these case studies and other examples to clarify ideas and provide examples of how researchers are approaching topics using clear causal thinking.

The collection and analysis of risk data are essential for the insurance-business model. The models for evaluating risk and predicting events that trigger insurance policies are based on knowledge derived from risk data.
The purpose of this project is to assess the scalability and flexibility of the software-based secure computing techniques in an insurance benchmarking scenario and to demonstrate the range of analytics capabilities they provide. These techniques offer provable technological guarantees that only authorized users can access the global models (fraud and loss models) based on the data of a network of collaborating organizations. The system relies on a fully distributed architecture without a centralized database, and implements advanced privacy-protection techniques based on multiparty homomorphic encryption, which makes it possible to efficiently compute machine-learning models on encrypted distributed data.

In communication systems, there are many tasks, like modulation recognition, for which Deep Neural Networks (DNNs) have obtained promising performance. However, these models have been shown to be susceptible to adversarial perturbations, namely imperceptible additive noise crafted to induce misclassification. This raises questions about the security but also the general trust in model predictions. In this project, we propose to use adversarial training, which consists of fine-tuning the model with adversarial perturbations, to increase the robustness of automatic modulation recognition (AMC) models. We show that current state-of-the-art models benefit from adversarial training, which mitigates the robustness issues for some families of modulations. We use adversarial perturbations to visualize the features learned, and we found that in robust models the signal symbols are shifted towards the nearest classes in constellation space, like maximum likelihood methods. This confirms that robust models not only are more secure, but also more interpretable, building their decisions on signal statistics that are relevant to modulation recognition.

This work aims at creating a Proof of Concept of storing and managing data on a blockchain. This work answers the following two use-cases: (i) compliant storage, transfer and access management of (personal) sensitive data and (ii) compliant cross-border or cross-jurisdiction data sharing.

DEDIS brings to the table a permissioned blockchain and distributed ledger using a fast catch up mechanism that allows for very fast processing of the requests, while staying secure. It also includes a novel approach to encryption and decryption, where no central point of failure can let the documents be published to outsiders (Calypso). Swiss Re brings to the table interesting use cases which will require DEDIS to extend Calypso to implement data location policies.

MedCo, developed in the LDS lab of professor Jean-Pierre Hubaux in collaboration with professor Bryan Ford’s DEDIS lab and the Lausanne University Hospital (CHUV), is the first operational system that makes sensitive medical-data available for research in a simple, privacy-conscious and secure way. It enables hundreds of clinical sites to collectively protect their data and to securely share them with investigators, without single points of failure. MedCo applies advanced privacy-enhancing techniques, such as: Multi-party homomorphic encryption, Secure distributed protocols and Differential privacy.

The DEDIS team created a first version of the onChain secrets implementation using its skipchain blockchain. This implementation allows a client to store encrypted documents on a public but permissioned blockchain and to change the access rights to those documents after they have been written to the blockchain. The first implementation has been extensively tested by ByzGen and is ready to be used in a PoC demo.
This project aims at increasing its performance and stability, and make it production-ready. Further, it will add a more realistic testing platform that will allow to check the validity of new functionality in a real-world setting and find regressions before they are pushed to the stable repository.

SafeAI aims to develop cyber-security solutions in the context of Artificial Intelligence (AI). With the advent of generative AI, it is possible to attack AI enhanced applications with targeted cyberattacks, and also to generate cyberattacks that are automated and enhanced via the use of AI. The main goal of SafeAI is the development of a software that enables automated generation of adversarial attacks and defences using AI.