A big convergence of language, vision, and multimodal pretraining is emerging. In this work, we introduce a general-purpose multimodal foundation model BEiT-3, which achieves state-of-the-art transfer performance on both vision and vision-language tasks. Specifically, we advance the big convergence from three aspects: backbone architecture, pretraining task, and model scaling up. We introduce Multiway Transformers for general-purpose modeling, where the modular architecture enables both deep fusion and modality-specific encoding. Based on the shared backbone, we perform masked "language" modeling on images (Imglish), texts (English), and image-text pairs ("parallel sentences") in a unified manner. Experimental results show that BEiT-3 obtains state-of-the-art performance on object detection (COCO), semantic segmentation (ADE20K), image classification (ImageNet), visual reasoning (NLVR2), visual question answering (VQAv2), image captioning (COCO), and cross-modal retrieval (Flickr30K, COCO).

Structural health monitoring (SHM) has been an active research area for the last three decades, and has accumulated a number of critical advances over that period, as can be seen in the literature. However, SHM is still facing challenges because of the paucity of damage-state data, operational and environmental fluctuations, repeatability issues, and changes in boundary conditions. These issues present as inconsistencies in the captured features and can have a huge impact on the practical implementation, but more critically, on the generalisation of the technology. Population-based SHM has been designed to address some of these concerns by modelling and transferring missing information using data collected from groups of similar structures. In this work, vibration data were collected from four healthy, nominally-identical, full-scale composite helicopter blades. Manufacturing differences (e.g., slight differences in geometry and/or material properties), among the blades presented as variability in their structural dynamics, which can be very problematic for SHM based on machine learning from vibration data. This work aims to address this variability by defining a general model for the frequency response functions of the blades, called a form, using mixtures of Gaussian processes.

Camera traps have revolutionized the animal research of many species that were previously nearly impossible to observe due to their habitat or behavior. They are cameras generally fixed to a tree that take a short sequence of images when triggered. Deep learning has the potential to overcome the workload to automate image classification according to taxon or empty images. However, a standard deep neural network classifier fails because animals often represent a small portion of the high-definition images. That is why we propose a workflow named Weakly Object Detection Faster-RCNN+FPN which suits this challenge. The model is weakly supervised because it requires only the animal taxon label per image but doesn't require any manual bounding box annotations. First, it automatically performs the weakly-supervised bounding box annotation using the motion from multiple frames. Then, it trains a Faster-RCNN+FPN model using this weak supervision. Experimental results have been obtained with two datasets from a Papua New Guinea and Missouri biodiversity monitoring campaign, then on an easily reproducible testbed.

In Federated Learning (FL) of click-through rate (CTR) prediction, users' data is not shared for privacy protection. The learning is performed by training locally on client devices and communicating only model changes to the server. There are two main challenges: (i) the client heterogeneity, making FL algorithms that use the weighted averaging to aggregate model updates from the clients have slow progress and unsatisfactory learning results; and (ii) the difficulty of tuning the server learning rate with trial-and-error methodology due to the big computation time and resources needed for each experiment. To address these challenges, we propose a simple online meta-learning method to learn a strategy of aggregating the model updates, which adaptively weighs the importance of the clients based on their attributes and adjust the step sizes of the update. We perform extensive evaluations on public datasets. Our method significantly outperforms the state-of-the-art in both the speed of convergence and the quality of the final learning results.

One of the most fundamental problems in machine learning is to compare examples: Given a pair of objects we want to return a value which indicates degree of (dis)similarity. Similarity is often task specific, and pre-defined distances can perform poorly, leading to work in metric learning. However, being able to learn a similarity-sensitive distance function also presupposes access to a rich, discriminative representation for the objects at hand. In this dissertation we present contributions towards both ends. In the first part of the thesis, assuming good representations for the data, we present a formulation for metric learning that makes a more direct attempt to optimize for the k-NN accuracy as compared to prior work. We also present extensions of this formulation to metric learning for kNN regression, asymmetric similarity learning and discriminative learning of Hamming distance. In the second part, we consider a situation where we are on a limited computational budget i.e. optimizing over a space of possible metrics would be infeasible, but access to a label aware distance metric is still desirable. We present a simple, and computationally inexpensive approach for estimating a well motivated metric that relies only on gradient estimates, discussing theoretical and experimental results. In the final part, we address representational issues, considering group equivariant convolutional neural networks (GCNNs). Equivariance to symmetry transformations is explicitly encoded in GCNNs; a classical CNN being the simplest example. In particular, we present a SO(3)-equivariant neural network architecture for spherical data, that operates entirely in Fourier space, while also providing a formalism for the design of fully Fourier neural networks that are equivariant to the action of any continuous compact group.

Many U.S. metropolitan cities are notorious for their severe shortage of parking spots. To this end, we present a proactive prediction-driven optimization framework to dynamically adjust parking prices. We use state-of-the-art deep learning technologies such as neural ordinary differential equations (NODEs) to design our future parking occupancy rate prediction model given historical occupancy rates and price information. Owing to the continuous and bijective characteristics of NODEs, in addition, we design a one-shot price optimization method given a pre-trained prediction model, which requires only one iteration to find the optimal solution. In other words, we optimize the price input to the pre-trained prediction model to achieve targeted occupancy rates in the parking blocks. We conduct experiments with the data collected in San Francisco and Seattle for years. Our prediction model shows the best accuracy in comparison with various temporal or spatio-temporal forecasting models. Our one-shot optimization method greatly outperforms other black-box and white-box search methods in terms of the search time and always returns the optimal price solution.

When an agent encounters a continual stream of new tasks in the lifelong learning setting, it leverages the knowledge it gained from the earlier tasks to help learn the new tasks better. In such a scenario, identifying an efficient knowledge representation becomes a challenging problem. Most research works propose to either store a subset of examples from the past tasks in a replay buffer, dedicate a separate set of parameters to each task or penalize excessive updates over parameters by introducing a regularization term. While existing methods employ the general task-agnostic stochastic gradient descent update rule, we propose a task-aware optimizer that adapts the learning rate based on the relatedness among tasks. We utilize the directions taken by the parameters during the updates by additively accumulating the gradients specific to each task. These task-based accumulated gradients act as a knowledge base that is maintained and updated throughout the stream. We empirically show that our proposed adaptive learning rate not only accounts for catastrophic forgetting but also exhibits knowledge transfer. We also show that our method performs better than several state-of-the-art methods in lifelong learning on complex datasets. Moreover, our method can also be combined with the existing methods and achieve substantial improvement in performance. Lifelong learning (LLL), also known as continual learning, is a setting where an agent continuously learns from data belonging to different tasks (Parisi et al., 2019). Here, the goal is to maximize performance on all the tasks arriving in a stream without replaying the entire datasets from past tasks (Riemer et al., 2018). Approaches proposed in this setting involve investigating the stability-plasticity dilemma (Mermillod et al., 2013) in different ways where stability refers to preventing the forgetting of past knowledge and plasticity refers to accumulating new knowledge by learning new tasks (Mermillod et al., 2013; Delange et al., 2021). 1 Unlike human beings, who can efficiently assess the correctness and applicability of the past knowledge (Chen & Liu, 2018), neural networks and other machine learning models often face various issues in this setting. Whenever data from a new task arrives, these models often tend to forget the previously obtained knowledge due to dependency on the input data distribution, limited capacity, diversity among tasks, etc.

The recent privacy leakage incidences and the more strict policy regulations demand a much higher standard of compliance for companies and mobile apps. However, such obligations also impose significant challenges on app developers for complying with these regulations that contain various perspectives, activities, and roles, especially for small companies and developers who are less experienced in this matter or with limited resources. To address these hurdles, we develop an automatic tool, NL2GDPR, which can generate policies from natural language descriptions from the developer while also ensuring the app's functionalities are compliant with General Data Protection Regulation (GDPR). NL2GDPR is developed by leveraging an information extraction tool, OIA (Open Information Annotation), developed by Sun et al. (2020); Wang et al. (2022b) from Baidu Cognitive Computing Lab. At the core, NL2GDPR is a privacy-centric information extraction model, appended with a GDPR policy finder and a policy generator. We perform a comprehensive study to grasp the challenges in extracting privacy-centric information and generating privacy policies, while exploiting optimizations for this specific task. With NL2GDPR, we can achieve 92.9%, 95.2%, and 98.4% accuracy in correctly identifying GDPR policies related to personal data storage, process, and share types, respectively. To the best of our knowledge, NL2GDPR is the first tool that allows a developer to automatically generate GDPR compliant policies, with only the need of entering the natural language for describing the app features. Note that other non-GDPR-related features might be integrated with the generated features to build a complex app.

Inverse design, the process of matching a device or process parameters to exhibit a desired performance, is applied in many disciplines ranging from material design over chemical processes and to engineering. Machine learning has emerged as a promising approach to overcome current limitations imposed by the dimensionality of the parameter space and multimodal parameter distributions. Most traditional optimization routines assume an invertible one-to-one mapping between the design parameters and the target performance. However, comparable or even identical performance may be realized by different designs, yielding a multimodal distribution of possible solutions to the inverse design problem which confuses the optimization algorithm. Here, we show how a generative modeling approach based on invertible neural networks can provide the full distribution of possible solutions to the inverse design problem and resolve the ambiguity of nanodevice inverse design problems featuring multimodal distributions. We implement a Conditional Invertible Neural Network (cINN) and apply it to a proof-of-principle nanophotonic problem, consisting in tailoring the transmission spectrum of a metallic film milled by subwavelength indentations. We compare our approach with the commonly used conditional Variational Autoencoder (cVAE) framework and show the superior flexibility and accuracy of the proposed cINNs when dealing with multimodal device distributions. Our work shows that invertible neural networks provide a valuable and versatile toolkit for advancing inverse design in nanoscience and nanotechnology.

Multi-view learning has progressed rapidly in recent years. Although many previous studies assume that each instance appears in all views, it is common in real-world applications for instances to be missing from some views, resulting in incomplete multi-view data. To tackle this problem, we propose a novel Latent Heterogeneous Graph Network (LHGN) for incomplete multi-view learning, which aims to use multiple incomplete views as fully as possible in a flexible manner. By learning a unified latent representation, a trade-off between consistency and complementarity among different views is implicitly realized. To explore the complex relationship between samples and latent representations, a neighborhood constraint and a view-existence constraint are proposed, for the first time, to construct a heterogeneous graph. Finally, to avoid any inconsistencies between training and test phase, a transductive learning technique is applied based on graph learning for classification tasks. Extensive experimental results on real-world datasets demonstrate the effectiveness of our model over existing state-of-the-art approaches.