Applications of artificial intelligence

Artificial intelligence (AI) has been used in applications to alleviate certain problems throughout industry and academia. AI, like electricity or computers, is a general purpose technology that has a multitude of applications. It has been used in fields of language translation, image recognition, credit scoring, e-commerce and other domains. Internet and e-commerce Recommendation systems Recommendation system A recommendation system predicts the "rating" or "preference" a user would give to an item. Recommendation systems are used in a variety of areas, such as generating playlists for video and music services, product recommendations for online stores, or content recommendations for social media platforms and open web content recommendation. Companies to use such systems include Netflix, Amazon and YouTube. Web feeds and posts Machine learning is also used in web feeds such as for determining which posts show up in social media feeds. Various types

Towards Verifiable, Generalizable and Efficient Robust Deep Neural Networks.

Chen Liu

In the last decade, deep neural networks have achieved tremendous success in many fields of machine learning.However, they are shown vulnerable against adversarial attacks: well-designed, yet imperceptible, perturbations can make the state-of-the-art deep neural networks output incorrect results.Understanding adversarial attacks and designing algorithms to make deep neural networks robust against these attacks are key steps to building reliable artificial intelligence in real-life applications.In this thesis, we will first formulate the robust learning problem.Based on the notations of empirical robustness and verified robustness, we design new algorithms to achieve both of these types of robustness.Specifically, we investigate the robust learning problem from the optimization perspectives.Compared with classic empirical risk minimization, we show the slow convergence and large generalization gap in robust learning.Our theoretical and numerical analysis indicates that these challenges arise, respectively, from non-smooth loss landscapes and model's fitting hard adversarial instances.Our insights shed some light on designing algorithms for mitigating these challenges.Robust learning has other challenges, such as large model capacity requirements and high computational complexity.To solve the model capacity issue, we combine robust learning with model compression.We design an algorithm to obtain sparse and binary neural networks and make it robust.To decrease the computational complexity, we accelerate the existing adversarial training algorithm and preserve its performance stability.In addition to making models robust, our research provides other benefits.Our methods demonstrate that robust models, compared with non-robust ones, usually utilize input features in a way more similar to the way human beings use them, hence the robust models are more interpretable.To obtain verified robustness, our methods indicate the geometric similarity of the decision boundaries near data points.Our approaches towards reliable artificial intelligence can not only render deep neural networks more robust in safety-critical applications but also make us better aware of how they work.

EPFL2022

A 3D indicator for guiding AI applications in the energy sector

Christophe Ballif, Marine Dominique C. Cauz, Laure-Emmanuelle Perret Aebi, Hugo James André Quest, Christian Rod, Alessandro Francesco Aldo Virtuani

The utilisation of Artificial Intelligence (AI) applications in the energy sector is gaining momentum, with increasingly intensive search for suitable, high-quality and trustworthy solutions that displayed promising results in research. The growing interest comes from decision makers of both the industry and policy domains, searching for applications to increase companies’ profitability, raise efficiency and facilitate the energy transition. This paper aims to provide a novel three-dimensional (3D) indicator for AI applications in the energy sector, based on their respective maturity level, regulatory risks and potential benefits. Case studies are used to exemplify the application of the 3D indicator, showcasing how the developed framework can be used to filter promising AI applications eligible for governmental funding or business development. In addition, the 3D indicator is used to rank AI applications considering different stakeholder preferences (risk-avoidance, profit-seeking, balanced). These results allow AI applications to be better categorised in the face of rapidly emerging national and intergovernmental AI strategies and regulations that constrain the use of AI applications in critical infrastructures.

2022

Towards Verifiable, Generalizable and Efficient Robust Deep Neural Networks.

Chen Liu

EPFL2022

Applications of artificial intelligence

Lévy processes and financial applications

Towards Verifiable, Generalizable and Efficient Robust Deep Neural Networks.

A 3D indicator for guiding AI applications in the energy sector

Towards Verifiable, Generalizable and Efficient Robust Deep Neural Networks.

Lévy processes and financial applications

A 3D indicator for guiding AI applications in the energy sector