Several domains increasingly rely on machine learning in their applications. The resulting heavy dependence on data has led to the emergence of various laws and regulations around data ethics and privacy and growing awareness of the need for privacy-preserving machine learning (ppML). Current ppML techniques utilize methods that are either purely based on cryptography, such as homomorphic encryption, or that introduce noise into the input, such as differential privacy. The main criticism given to those techniques is the fact that they either are too slow or they trade off a model's performance for improved confidentiality. To address this performance reduction, we aim to leverage robust representation learning as a way of encoding our data while optimising the privacy-utility trade-off. Our method centers on training autoencoders in a multi-objective manner and then concatenating the latent and learned features from the encoding part as the encoded form of our data. Such a deep learning-powered encoding can then safely be sent to a third party for intensive training and hyperparameter tuning. With our proposed framework, we can share our data and use third party tools without being under the threat of revealing its original form. We empirically validate our results on unimodal and multimodal settings, the latter following a vertical splitting system and show improved performance over state-of-the-art.

Accurate particle identification is an ongoing task in the European organization for nuclear research, known as CERN where the challenge remains that targeted particles/events represent tiny minorities in front of the overwhelming presence of common particles such as protons. This paper presents a directed undersampling using an active learning method named DUAL to handle the high imbalance problem present in the particle identification dataset. The proposed approach was used to reduce the training set size while maintaining classifiers' performance. Compared against various imbalance learning approaches, the experimental results show that using DUAL as a data reduction technique with a random forest classifier enhances classification performance in terms of Macro- $$F_1$$ score and decreases the training time needed to train the models, which is very relevant while dealing with large-scale datasets. Despite being experimented only with particle identification dataset, we believe that DUAL could be adopted as a generic method for multi-class imbalanced classification problems with big data scale difficulties.

In this study, we propose feature extraction for multimodal meme classification using Deep Learning approaches. A meme is usually a photo or video with text shared by the young generation on social media platforms that expresses a culturally relevant idea. Since they are an efficient way to express emotions and feelings, a good classifier that can classify the sentiment behind the meme is important. To make the learning process more efficient, reduce the likelihood of overfitting, and improve the generalizability of the model, one needs a good approach for joint feature extraction from all modalities. In this work, we proposed to use different multimodal neural network approaches for multimodal feature extraction and use the extracted features to train a classifier to identify the sentiment in a meme.