Current methods for sequence prediction often fail to account for higher-ordercontinuity. This results in the prediction of sequences that might be continuous but not physically viable as I investigate higher-order smoothness in terms of velocity and acceleration. Hence, I propose a Yuksel Spline-based model that is not only capable of predicting curves that are guaranteed to be C^2 continuous but, also efficient to compute as well. Characteristic properties of the models are demonstrated over toy examples and sequence prediction tasks.

The advancement of cloud technology has impacted society positively in a number of ways, but it has also led to an increase in threats that target private information available on cloud systems. Intrusion prevention systems play a crucial role in protecting cloud systems from such threats. In this thesis, an intrusion prevention approach todetect and prevent such threats in real-time is proposed. This approach is designed for network-based intrusion prevention systems and leverages the power of supervised machine learning with Extreme Gradient Boosting (XGBoost) and Long Short-Term Memory (LSTM) algorithms, to analyze the flow of each packet that is sent to a cloud system through the network. The innovations of this thesis include developing a custom LSTM architecture, using this architecture to train a LSTM model to identify attacks and using TCP reset functionality to prevent attacks for cloud systems. The aim of this thesis is to provide a framework for an Intrusion Prevention System. Based on simulations and experimental results with the NF-UQ-NIDS-v2 dataset, the proposed system is accurate, fast, scalable and has a low rate of false positives, making it suitable for real world applications.

The proposed research is motivated by the colon cancer bio-marker study, which recruited case (or colon cancer) and healthy control samples and quantified their large number of candidate bio-markers using a high-throughput technology, called nucleicacid-programmable protein array (NAPPA). The study aimed to identify a panel of biomarkers to accurately distinguish between the cases and controls. A major challenge in analyzing this study was the bio-marker heterogeneity, where bio-marker responses differ from sample to sample. The goal of this research is to improve prediction accuracy for motivating or similar studies. Most machine learning (ML) algorithms, developed under the one-size-fits-all strategy, were not able to analyze the above-mentioned heterogeneous data. Failing to capture the individuality of each subject, several standard ML algorithms tested against this dataset performed poorly resulting in 55-61% accuracy. Alternatively, the proposed personalized ML (PML) strategy aims at tailoring the optimal ML models for each subject according to their individual characteristics yielding best highest accuracy of 72%.

With the rise in social media usage and rapid communication, the proliferation of misinformation and fake news has become a pressing concern. The detection of multimodal fake news requires careful consideration of both image and textual semantics with proper alignment of the embedding space. Automated fake news detection has gained significant attention in recent years. Existing research has focused on either capturing cross-modal inconsistency information or leveraging the complementary information within image-text pairs. However, the potential of powerful cross-modal contrastive learning methods and effective modality mixing remains an open-ended question. The thesis proposes a novel two-leg single-tower architecture equipped with self-attention mechanisms and custom contrastive loss to efficiently aggregate multimodal features. Furthermore, pretraining and fine-tuning are employed on the custom transformer model to classify fake news across the popular Twitter multimodal fake news dataset. The experimental results demonstrate the efficacy and robustness of the proposed approach, offering promising advancements in multimodal fake news detection research.

Multimodal reasoning is one of the most interesting research fields because of the ability to interact with systems and the explainability of the models' behavior. Traditional multimodal research problems do not focus on complex commonsense reasoning (such as physical interactions). Although real-world objects have physical properties associated with them, many of these properties (such as mass and coefficient of friction) are not captured directly by the imaging pipeline. Videos often capture objects, their motion, and the interactions between different objects. However, these properties can be estimated by utilizing cues from relative object motion and the dynamics introduced by collisions. This thesis introduces a new video question-answering task for reasoning about the implicit physical properties of objects in a scene, from videos. For this task, I introduce a dataset -- CRIPP-VQA (Counterfactual Reasoning about Implicit Physical Properties - Video Question Answering), which contains videos of objects in motion, annotated with hypothetical/counterfactual questions about the effect of actions (such as removing, adding, or replacing objects), questions about planning (choosing actions to perform to reach a particular goal), as well as descriptive questions about the visible properties of objects. Further, I benchmark the performance of existing video question-answering models on two test settings of CRIPP-VQA: i.i.d. and an out-of-distribution setting which contains objects with values of mass, coefficient of friction, and initial velocities that are not seen in the training distribution. Experiments reveal a surprising and significant performance gap in terms of answering questions about implicit properties (the focus of this thesis) and explicit properties (the focus of prior work) of objects.

The study tested the parameterized neural ordinary differential equation (PNODE) framework with a physical system exhibiting only advective phenomenon. Existing deep learning methods have difficulty learning multiple dynamic, continuous time processes. PNODE encodes the input data and initial parameter into a set of reduced states within the latent space. Then the reduced states are fitted to a system of ordinary differential equations. The outputs from the model are then decoded back to the data space for a desired input parameter and time. The application of the PNODE formalism to different types of physical systems is important to test the methods robustness. The linear advection data was generated through a high-fidelity numerical tool for multiple velocity parameters. The PNODE code was modified for the advection dataset, whose temporal domain and spatial discretization varied from the original study configuration. The L2 norm between the reconstruction and surrogate model and the reconstruction plots were used to analyze the PNODE model performance. The model reconstructions presented mixed results. For a temporal domain of 20-time units, where multiple advection cycles were completed for each advection speed, the reconstructions did not agree with the surrogate model. For a reduced temporal domain of 5-time units, the reconstructions and surrogate models were in close agreement. Near the end of the temporal domain, deviations occurred likely resulting from the accumulation of numerical errors. Note, over the 5-time units, smaller advection speed parameters were unable to complete a cycle. The behavior for the 20-time units highlighted potential issues with imbalanced datasets and repeated features. The 5-time unit model illustrates PNODEs adaptability to this class of problems when the dataset is better posed.