Detecting Latent Training Needs Using Large Datasets

In today's world, there is no shortage of disruptors acting on various professional domains. The Fourth Industrial Revolution, with its AI-driven and automation-focused technologies, has fundamentally changed many domains -- particularly the Information and Communication Technologies (ICT) domain -- and continues to do so. The COVID-19 pandemic and its ensuing lockdowns have resulted in dramatic change in the workplace. These disruptors are fast-acting and rapidly change the skills landscape of many professional domains, introducing new skills and rendering old ones obsolete. In such a situation, it is imperative for educational institutions to keep their curricula updated in order to (re)train the workforce with the right set of skills for the new economy. However, most of the methodologies commonly used for analyzing training needs and effecting curricular change were developed during a time when the pace of change was slower, and thus have trouble quickly and effectively responding to today's rapid changes. As a result, the development of new approaches for the analysis of training needs is of utmost importance.The present dissertation is the result of an effort to develop such new approaches, using different types of data and approaching the problem from different perspectives. These approaches use big data methods on data sources that are large, pre-existing, and continually updated in order to identify the skill needs of the present and/or predict the skill needs of the future for various professions. Focusing on the ICT domain but also making forays into vocational domains, the present work examines the usefulness and feasibility of methodologies such as the estimation of topical difficulty, the prediction of the appearance of new skills, and the prediction of emerging skills (less popular skills expected to experience a surge in hiring demand) for the identification of training needs. The findings of this dissertation are twofold. Firstly, they demonstrate the pivotal role that decentralized educational platforms such as Stack Overflow (an online Q&A platform for ICT) and Udemy (a MOOC platform where anyone can create a course) play in lifelong learning, essentially making the case for a policy of sponsorship and promotion of such platforms in other professional domains. Secondly and more importantly, the results showcase the feasibility of predicting the emerging skills of the near future. The successful emerging skills methodology can therefore be used to provide valuable insights to training providers at a moment's notice, giving them time to prepare for the skills landscape of the future in advance. Thus, AI can be used to help solve the challenges it creates for educational institutions and the workforce. Future work, both in the form of policymaking and in the form of research, can help capitalize and expand on the results of this dissertation, furthering the cause of rapid curricular change.

Optimization methods for collaborative learning

Sai Praneeth Reddy Karimireddy

A traditional machine learning pipeline involves collecting massive amounts of data centrally on a server and training models to fit the data. However, increasing concerns about the privacy and security of user's data, combined with the sheer growth in the data sizes has incentivized looking beyond such traditional centralized approaches. Collaborative learning (which encompasses distributed, federated, and decentralized learning) proposes instead for a network of data holders to collaborate together to train models without transmitting any data. This new paradigm minimizes data exposure, but inherently faces some fundamental challenges. In this thesis, we bring to bear the framework of stochastic optimization to formalize and develop new algorithms for these challenges. This serves not only to develop novel solutions, but also to test the utility of the optimization lens in modern deep learning.We study three fundamental problems. Firstly, collaborative training replaces a one-time transmission of raw data with repeated rounds of communicating partially trained models. However, this quickly runs against bandwidth constraints when dealing with large models. We propose to solve this bandwidth constraint using compressed communication. Next, collaborative training leverages the computation power of the data holders directly. However, this is not as reliable as using a data center with only a subset of them available at any given time. Thus, we require new algorithms which can efficiently utilize unreliable local computation of the data holders. Finally, collaborative training allows any data holder to participate in the training process, without being able to inspect their data or local computation. This may potentially open the system to malicious or faulty agents who seek to derail the training. We develop algorithms with Byzantine robustness which are guaranteed to be resilient to such attackers.

EPFL2021

Neural Scene Decomposition for Multi-Person Motion Capture

Victor Constantin, Pascal Fua, Isinsu Katircioglu, Helge Jochen Rhodin, Mathieu Salzmann

Learning general image representations has proven key to the success of many computer vision tasks. For example, many approaches to image understanding problems rely on deep networks that were initially trained on ImageNet, mostly because the learned features are a valuable starting point to learn from limited labeled data. However, when it comes to 3D motion capture of multiple people, these features are only of limited use. In this paper, we therefore propose an approach to learning features that are useful for this purpose. To this end, we introduce a self-supervised approach to learning what we call a neural scene decomposition (NSD) that can be exploited for 3D pose estimation. NSD comprises three layers of abstraction to represent human subjects: spatial layout in terms of bounding-boxes and relative depth; a 2D shape representation in terms of an instance segmentation mask; and subject-specific appearance and 3D pose information. By exploiting self-supervision coming from multiview data, our NSD model can be trained end-to-end without any 2D or 3D supervision. In contrast to previous approaches, it works for multiple persons and full-frame images. Because it encodes 3D geometry, NSD can then be effectively leveraged to train a 3D pose estimation network from small amounts of annotated data. Our code and newly introduced boxing dataset is available at github.com and cvlab.epfl.ch.

IEEE2019

Bridging the gap between model-driven and data-driven methods in the era of Big Data

Gael Lederrey

Data-driven and model-driven methodologies can be regarded as competitive fields since they tackle similar problems such as prediction. However, these two fields can learn from each other to improve themselves. Indeed, data-driven methodologies have been developed to use advanced methodologies based on Big Data technologies. On the other hand, model-driven methodologies concentrate on developing mathematical models based on theory and expert knowledge to allow for interpretability and control. Through three main contributions, this thesis aims to bridge the gap between these two fields by using their strengths and applying them to its counterpart.Discrete Choice Models (DCMs) have shown tremendous success in many fields, such as transportation. However, they have not evolved to tackle the growing amount of available data. On the other hand, Machine Learning (ML) researchers have developed optimization algorithms to efficiently estimate complex models on large datasets. Similarly, faster estimation of DCMs on larger datasets would improve the efficiency of modelers as well as enable new research axes. Thus, we take inspiration from the large body of existing research in efficient parameter estimation with extensive data and large numbers of parameters in deep learning and apply it to DCMs. The first chapter of this thesis introduces the HAMABS algorithm, which combines three fundamental principles to enable faster parameter estimation of DCMs (20x speedup compared to standard estimation) without compromising the precision of the parameter estimates.Collecting large amounts of data can be cumbersome and costly, even in the era of Big Data. For example, ML researchers in Computer Vision have been developing generative deep learning models to augment datasets. DCM researchers face similar issues with tabular data, e.g. travel surveys. In addition, if the collection process is not performed correctly, these datasets can contain bias, lack consistency, or be unrepresentative of the actual population. The second chapter of this thesis introduces the DATGAN, a Generative Adversarial Network (GAN) integrating expert knowledge to control the generation process. This new architecture allows modelers to generate controlled and representative synthetic data, outperforming similar state-of-the-art generative models. Finally, researchers are increasingly developing fully disaggregate agent-based simulation models, which use detailed synthetic populations to generate aggregate passenger flows. However, detailed disaggregate socioeconomic data is usually expensive to collect and heavily restricted in terms of access and usage. As such, synthetic populations are typically either drawn randomly from aggregate level control totals, limiting their quality, or tightly controlled, limiting their application and usefulness. To combat this, the third chapter extends the DATGAN methodology to generate highly detailed and consistent synthetic populations from small sample data. First, ciDATGAN learns to generate the variables in a low-sample highly detailed dataset, e.g. household travel survey. It then completes a high-sample dataset with few variables, e.g. microdata census, by generating the previously learned variables. The results show that this methodology can correct for bias and may enable the transfer of synthetic populations to new areas/contexts.