学习推迟（L2D）框架有可能使AI系统更安全。对于给定的输入，如果人类比模型更有可能采取正确的行动，则系统可以将决定推迟给人类。我们研究L2D系统的校准，研究它们输出的概率是否合理。我们发现Mozannar＆Sontag（2020）多类框架没有针对专家正确性进行校准。此外，由于其参数化是为此目的而退化的，因此甚至不能保证产生有效的概率。我们提出了一个基于单VS-ALL分类器的L2D系统，该系统能够产生专家正确性的校准概率。此外，我们的损失功能也是多类L2D的一致替代，例如Mozannar＆Sontag（2020）。我们的实验验证了我们的系统校准不仅是我们的系统校准，而且这种好处无需准确。我们的模型的准确性始终可与Mozannar＆Sontag（2020）模型的模型相当（通常是优越），从仇恨言语检测到星系分类到诊断皮肤病变的任务。

由于需要将靠近用户的所有处理和解决隐私问题需要，人工智能现在在智能手机行业中占据了智能手机行业的中心阶段。若干AI应用程序使用的卷积神经网络（CNNS）是高度资源和计算密集型。虽然新一代智能手机具有启用AI的芯片，但最小的内存和能量利用率对于许多应用程序在智能手机上同时运行。鉴于此，通过将处理的一部分卸载到云服务器的一部分来优化智能手机上的工作负载是一个重要的研究方向。在本文中，我们通过制定优化端到端延迟，内存利用率和能量消耗的多目标优化问题来分析智能手机和云服务器之间分离CNN的可行性。我们设计SmartSplit，一种基于决策分析的遗传算法来解决优化问题。我们使用多个CNN模型运行的实验显示，在智能手机和云服务器之间拆分CNN是可行的。与其他最先进的方法相比，SmartSplit的方法，SmartSplit更好。

联合学习偏离“将数据发送到模型”的规范“向数据发送模型”。当在边缘生态系统中使用时，许多异构边缘设备通过不同的方式收集数据并通过不同的网络信道连接参与培训过程。由于设备故障或网络问题，这种生态系统中的边缘设备的失败很可能。在本文中，我们首先分析边缘设备数量对FL模型的影响，并提供一种选择有助于该模型的最佳设备的策略。我们观察所选设备失败并提供缓解策略以确保强大的联合学习技术的影响。

A recent trend in deep learning research features the application of graph neural networks for mesh-based continuum mechanics simulations. Most of these frameworks operate on graphs in which each edge connects two nodes. Inspired by the data connectivity in the finite element method, we connect the nodes by elements rather than edges, effectively forming a hypergraph. We implement a message-passing network on such a node-element hypergraph and explore the capability of the network for the modeling of fluid flow. The network is tested on two common benchmark problems, namely the fluid flow around a circular cylinder and airfoil configurations. The results show that such a message-passing network defined on the node-element hypergraph is able to generate more stable and accurate temporal roll-out predictions compared to the baseline generalized message-passing network defined on a normal graph. Along with adjustments in activation function and training loss, we expect this work to set a new strong baseline for future explorations of mesh-based fluid simulations with graph neural networks.

Mixup is a popular data augmentation technique for training deep neural networks where additional samples are generated by linearly interpolating pairs of inputs and their labels. This technique is known to improve the generalization performance in many learning paradigms and applications. In this work, we first analyze Mixup and show that it implicitly regularizes infinitely many directional derivatives of all orders. We then propose a new method to improve Mixup based on the novel insight. To demonstrate the effectiveness of the proposed method, we conduct experiments across various domains such as images, tabular data, speech, and graphs. Our results show that the proposed method improves Mixup across various datasets using a variety of architectures, for instance, exhibiting an improvement over Mixup by 0.8% in ImageNet top-1 accuracy.

In multi-agent systems with large number of agents, typically the contribution of each agent to the value of other agents is minimal (e.g., aggregation systems such as Uber, Deliveroo). In this paper, we consider such multi-agent systems where each agent is self-interested and takes a sequence of decisions and represent them as a Stochastic Non-atomic Congestion Game (SNCG). We derive key properties for equilibrium solutions in SNCG model with non-atomic and also nearly non-atomic agents. With those key equilibrium properties, we provide a novel Multi-Agent Reinforcement Learning (MARL) mechanism that minimizes variance across values of agents in the same state. To demonstrate the utility of this new mechanism, we provide detailed results on a real-world taxi dataset and also a generic simulator for aggregation systems. We show that our approach reduces the variance in revenues earned by taxi drivers, while still providing higher joint revenues than leading approaches.

This thesis considers sequential decision problems, where the loss/reward incurred by selecting an action may not be inferred from observed feedback. A major part of this thesis focuses on the unsupervised sequential selection problem, where one can not infer the loss incurred for selecting an action from observed feedback. We also introduce a new setup named Censored Semi Bandits, where the loss incurred for selecting an action can be observed under certain conditions. Finally, we study the channel selection problem in the communication networks, where the reward for an action is only observed when no other player selects that action to play in the round. These problems find applications in many fields like healthcare, crowd-sourcing, security, adaptive resource allocation, among many others. This thesis aims to address the above-described sequential decision problems by exploiting specific structures these problems exhibit. We develop provably optimal algorithms for each of these setups with weak feedback and validate their empirical performance on different problem instances derived from synthetic and real datasets.

This thesis considers sequential decision problems, where the loss/reward incurred by selecting an action may not be inferred from observed feedback. A major part of this thesis focuses on the unsupervised sequential selection problem, where one can not infer the loss incurred for selecting an action from observed feedback. We also introduce a new setup named Censored Semi Bandits, where the loss incurred for selecting an action can be observed under certain conditions. Finally, we study the channel selection problem in the communication networks, where the reward for an action is only observed when no other player selects that action to play in the round. These problems find applications in many fields like healthcare, crowd-sourcing, security, adaptive resource allocation, among many others. This thesis aims to address the above-described sequential decision problems by exploiting specific structures these problems exhibit. We develop provably optimal algorithms for each of these setups with weak feedback and validate their empirical performance on different problem instances derived from synthetic and real datasets.

Masked Language Modeling (MLM) has proven to be an essential component of Vision-Language (VL) pretraining. To implement MLM, the researcher must make two design choices: the masking strategy, which determines which tokens to mask, and the masking rate, which determines how many tokens to mask. Previous work has focused primarily on the masking strategy while setting the masking rate at a default of 15\%. In this paper, we show that increasing this masking rate improves downstream performance while simultaneously reducing performance gap among different masking strategies, rendering the uniform masking strategy competitive to other more complex ones. Surprisingly, we also discover that increasing the masking rate leads to gains in Image-Text Matching (ITM) tasks, suggesting that the role of MLM goes beyond language modeling in VL pretraining.

Artificial Intelligence (AI) systems have been increasingly used to make decision-making processes faster, more accurate, and more efficient. However, such systems are also at constant risk of being attacked. While the majority of attacks targeting AI-based applications aim to manipulate classifiers or training data and alter the output of an AI model, recently proposed Sponge Attacks against AI models aim to impede the classifier's execution by consuming substantial resources. In this work, we propose \textit{Dual Denial of Decision (DDoD) attacks against collaborative Human-AI teams}. We discuss how such attacks aim to deplete \textit{both computational and human} resources, and significantly impair decision-making capabilities. We describe DDoD on human and computational resources and present potential risk scenarios in a series of exemplary domains.