This research combines Knowledge Distillation (KD) and Mixture of Experts (MoE) to develop modular, efficient multilingual language models. Key objectives include evaluating adaptive versus fixed alpha methods in KD and comparing modular MoE architectures for handling multi-domain inputs and preventing catastrophic forgetting. KD compresses large language models (LLMs) into smaller, efficient models, while MoE enhances modularity with specialized tasks. Experiments showed similar performance for both KD methods, with marginal improvements from adaptive alpha. A combined loss approach provided more stable learning. The router, trained to classify input sequences into English, French, German, or Python, achieved 99.95% precision, recall, and F1 score, with Logistic Regression being the most effective classifier. Evaluations of modular MoE architectures revealed that Pre-trained Language Experts (PLE) and Joint Expert Embedding Training (JEET) performed similarly, while the MoE with Common Expert (MoE-CE) setup showed slightly lower performance. Including a common expert in MoE-CE improved its performance. Studies on catastrophic forgetting indicated that sequential training led to significant forgetting, while single-session training with balanced batches and the MoE approach mitigated this issue. The MoE architecture preserved knowledge across multiple languages effectively. The research contributes open-sourced resources including the dataset (https://zenodo.org/doi/10.5281/zenodo.12677631), a balanced dataset creation tool (https://github.com/padas-lab-de/multi-language-dataset-creator), and the research codebase (https://github.com/ModMaamari/mixture-modular-experts).

[PhD thesis of FCP.] Nowadays, genetics studies large amounts of very diverse variables. Mathematical statistics has evolved in parallel to its applications, with much recent interest high-dimensional settings. In the genetics of human common disease, a number of relevant problems can be formulated as tests of independence. We show how defining adequate premetric structures on the support spaces of the genetic data allows for novel approaches to such testing. This yields a solid theoretical framework, which reflects the underlying biology, and allows for computationally-efficient implementations. For each problem, we provide mathematical results, simulations and the application to real data.

This paper presents a natural language processing (NLP) approach to the problem of thoroughly comprehending song lyrics, with particular attention on genre classification, view-based success prediction, and approximate release year. Our tests provide promising results with 65\% accuracy in genre classification and 79\% accuracy in success prediction, leveraging a DistilBERT model for genre classification and BERT embeddings for release year prediction. Support Vector Machines outperformed other models in predicting the release year, achieving the lowest root mean squared error (RMSE) of 14.18. Our study offers insights that have the potential to revolutionize our relationship with music by addressing the shortcomings of current approaches in properly understanding the emotional intricacies of song lyrics.

Meta-learning is a practical learning paradigm to transfer skills across tasks from a few examples. Nevertheless, the existence of task distribution shifts tends to weaken meta-learners' generalization capability, particularly when the task distribution is naively hand-crafted or based on simple priors that fail to cover typical scenarios sufficiently. Here, we consider explicitly generative modeling task distributions placed over task identifiers and propose robustifying fast adaptation from adversarial training. Our approach, which can be interpreted as a model of a Stackelberg game, not only uncovers the task structure during problem-solving from an explicit generative model but also theoretically increases the adaptation robustness in worst cases. This work has practical implications, particularly in dealing with task distribution shifts in meta-learning, and contributes to theoretical insights in the field. Our method demonstrates its robustness in the presence of task subpopulation shifts and improved performance over SOTA baselines in extensive experiments. The project is available at https://sites.google.com/view/ar-metalearn.

Solar energy is one of the most promising renewable energy resources. Forecasting photovoltaic power generation is an important way to increase photovoltaic penetration. However, the task of photovoltaic forecasting is complicated due to its property of uncertainty, especially in specific regions during the foggy winter. This paper proposes a novel model to accomplish the problem. A developed entropy is created to qualify the uncertainty during the foggy winter. The clustering method and modified retention network are applied to reduce complexity and forecast, respectively. We adopt an optimization to optimize the hyperparameters. Results are validated from the multivariate forecasting model using the dataset from a photovoltaic power station in Jiangsu Province, China. Experiments show that the proposed model improves the forecasting accuracy compared to various models during the foggy winter.

Graph data in real-world scenarios undergo rapid and frequent changes, making it challenging for existing graph models to effectively handle the continuous influx of new data and accommodate data withdrawal requests. The approach to frequently retraining graph models is resource intensive and impractical. To address this pressing challenge, this paper introduces a new concept of graph memory learning. Its core idea is to enable a graph model to selectively remember new knowledge but forget old knowledge. Building on this approach, the paper presents a novel graph memory learning framework - Brain-inspired Graph Memory Learning (BGML), inspired by brain network dynamics and function-structure coupling strategies. BGML incorporates a multi-granular hierarchical progressive learning mechanism rooted in feature graph grain learning to mitigate potential conflict between memorization and forgetting in graph memory learning. This mechanism allows for a comprehensive and multi-level perception of local details within evolving graphs. In addition, to tackle the issue of unreliable structures in newly added incremental information, the paper introduces an information self-assessment ownership mechanism. This mechanism not only facilitates the propagation of incremental information within the model but also effectively preserves the integrity of past experiences. We design five types of graph memory learning tasks: regular, memory, unlearning, data-incremental, and class-incremental to evaluate BGML. Its excellent performance is confirmed through extensive experiments on multiple real-world node classification datasets.

Deep Neural Networks (DNNs) are vulnerable to backdoor poisoning attacks, with most research focusing on digital triggers, special patterns digitally added to test-time inputs to induce targeted misclassification. In contrast, physical triggers, which are natural objects within a physical scene, have emerged as a desirable alternative since they enable real-time backdoor activations without digital manipulation. However, current physical attacks require that poisoned inputs have incorrect labels, making them easily detectable upon human inspection. In this paper, we collect a facial dataset of 21,238 images with 7 common accessories as triggers and use it to study the threat of clean-label backdoor attacks in the physical world. Our study reveals two findings. First, the success of physical attacks depends on the poisoning algorithm, physical trigger, and the pair of source-target classes. Second, although clean-label poisoned samples preserve ground-truth labels, their perceptual quality could be seriously degraded due to conspicuous artifacts in the images. Such samples are also vulnerable to statistical filtering methods because they deviate from the distribution of clean samples in the feature space. To address these issues, we propose replacing the standard $\ell_\infty$ regularization with a novel pixel regularization and feature regularization that could enhance the imperceptibility of poisoned samples without compromising attack performance. Our study highlights accidental backdoor activations as a key limitation of clean-label physical backdoor attacks. This happens when unintended objects or classes accidentally cause the model to misclassify as the target class.

Capturing the inter-dependencies among multiple types of clinicallycritical Time-to-event prediction (also known as survival analysis) investigates events is critical not only to accurate future event prediction, the distribution of time duration until the event of interest but also to better treatment planning. In this work, we propose a happens in the presence of event censorship. In the healthcare domain, deep latent state-space generative model to capture the interactions it is an essential tool for modeling the risks of critical medical among different types of correlated clinical events (e.g., kidney events and capturing of the relationship between the co-variants failure, mortality) by explicitly modeling the temporal dynamics and the risks [9]. of patients' latent states. Based on these learned patient states, we Recently, machine learning methods have been applied to timeto-event further develop a new general discrete-time formulation of the hazard predictions to provide flexible modeling of the time distribution rate function to estimate the survival distribution of patients [6, 19, 22], and capture the nonlinear relationship between with significantly improved accuracy.

The paradigm of'pretraining' from a set of relevant auxiliary tasks and then'finetuning' on a target task has been successfully applied in many different domains. However, when the auxiliary tasks are abundant, with complex relationships to the target task, using domain knowledge or searching over all possible pretraining setups is inefficient and suboptimal. To address this challenge, we propose a method to automatically select from a large set of auxiliary tasks, which yields a representation most useful to the target task. In particular, we develop an efficient algorithm that uses automatic auxiliary task selection within a nested-loop metalearning process. We have applied this algorithm to the task of clinical outcome predictions in electronic medical records, learning from a large number of selfsupervised tasks related to forecasting patient trajectories. Experiments on a real clinical dataset demonstrate the superior predictive performance of our method compared to direct supervised learning, naive pretraining and simple multitask learning, in particular in low-data scenarios when the primary task has very few examples. With detailed ablation analysis, we further show that the selection rules are interpretable and able to generalize to unseen target tasks with new data.

Personalized Federated Learning (FL) algorithms collaboratively train customized models for each client, enhancing the accuracy of the learned models on the client's local data (e.g., by clustering similar clients, or by fine-tuning models locally). In this paper, we investigate the impact of such personalization techniques on the group fairness of the learned models, and show that personalization can also lead to improved (local) fairness as an unintended benefit. We begin by illustrating these benefits of personalization through numerical experiments comparing two classes of personalized FL algorithms (clustering and fine-tuning) against a baseline FedAvg algorithm, elaborating on the reasons behind improved fairness using personalized FL, and then providing analytical support. Motivated by these, we further propose a new, Fairness-aware Federated Clustering Algorithm, Fair-FCA, in which clients can be clustered to obtain a (tuneable) fairness-accuracy tradeoff. Through numerical experiments, we demonstrate the ability of Fair-FCA to strike a balance between accuracy and fairness at the client level.