Streaming video is an ubiquitous service used by billions of people every day. Offering the best-in-class Quality of Experience (QoE) is a challenge to Mobile Network Operators (MNOs) which requires them to identify the obstacles to overcome. To properly achieve this, MNOs require efficient monitoring able to maintain this QoE at optimum levels. Machine Learning (ML) models play an essential role in this task. This paper introduces PenFFR, a novel approach based on Functional Data Analysis (FDA), to identify the underlying functions between the QoE and the features that characterize the streaming videos thus enabling to predict future values. The experimental results show that the performance of the prediction is on par for one metric with Deep Learning-based methods and even better for another metric. To facilitate the understanding of our method and enable hands-on application, the code is made available [1].

Authors publish documents in a dynamic manner. Their topic of interest and writing style might shift over time. Tasks such as author classification, author identification or link prediction are difficult to solve in such complex data settings. We propose a new representation learning model, DGEA (for Dynamic Gaussian Embedding of Authors), that is more suited to solve these tasks by capturing this temporal evolution. We formulate a general embedding framework: author representation at time t is a Gaussian distribution that leverages pre-trained document vectors, and that depends on the publications observed until t. The representations should retain some form of multi-topic information and temporal smoothness. We propose two models that fit into this framework. The first one, K-DGEA, uses a first order Markov model optimized with an Expectation Maximization Algorithm with Kalman Equations. The second, R-DGEA, makes use of a Recurrent Neural Network to model the time dependence. We evaluate our method on several quantitative tasks: author identification, classification, and co-authorship prediction, on two datasets written in English. In addition, our model is language agnostic since it only requires pre-trained document embeddings. It outperforms existing baselines by up to 18% on an author classification task on a news articles dataset.

Gaussian Embedding of Linked Documents (GELD) is a new method that embeds linked documents (e.g., citation networks) onto a pretrained semantic space (e.g., a set of word embeddings). We formulate the problem in such a way that we model each document as a Gaussian distribution in the word vector space. We design a generative model that combines both words and links in a consistent way. Leveraging the variance of a document allows us to model the uncertainty related to word and link generation. In most cases, our method outperforms state-of-the-art methods when using our document vectors as features for usual downstream tasks. In particular, GELD achieves better accuracy in classification and link prediction on Cora and Dblp. In addition, we demonstrate qualitatively the convenience of several properties of our method. We provide the implementation of GELD and the evaluation datasets to the community (https://github.com/AntoineGourru/DNEmbedding).

In order to help the network maintainers with the daily diagnosis and optimization tasks, a supervised model for mobile anomalies prevention is proposed. The objective is to detect future malfunctions of a set of cells, by only observing key performance indicators that are considered as functional data. Thus, by alerting the engineers as well as self-organizing networks, mobile operators can be saved from a certain performance degradation. The model has proven its efficiency with an application on real data that aims to detect capacity degradation, accessibility and call drops anomalies for LTE networks.