We are currently facing unprecedented cyber warfare with the rapid evolution of tactics, increasing asymmetry of intelligence, and the growing accessibility of hacking tools. In this landscape, cyber deception emerges as a critical component of our defense strategy against increasingly sophisticated attacks. This chapter aims to highlight the pivotal role of game-theoretic models and foundation models (FMs) in analyzing, designing, and implementing cyber deception tactics. Game models (GMs) serve as a foundational framework for modeling diverse adversarial interactions, allowing us to encapsulate both adversarial knowledge and domain-specific insights. Meanwhile, FMs serve as the building blocks for creating tailored machine learning models suited to given applications. By leveraging the synergy between GMs and FMs, we can advance proactive and automated cyber defense mechanisms by not only securing our networks against attacks but also enhancing their resilience against well-planned operations. This chapter discusses the games at the tactical, operational, and strategic levels of warfare, delves into the symbiotic relationship between these methodologies, and explores relevant applications where such a framework can make a substantial impact in cybersecurity. The chapter discusses the promising direction of the multi-agent neurosymbolic conjectural learning (MANSCOL), which allows the defender to predict adversarial behaviors, design adaptive defensive deception tactics, and synthesize knowledge for the operational level synthesis and adaptation. FMs serve as pivotal tools across various functions for MANSCOL, including reinforcement learning, knowledge assimilation, formation of conjectures, and contextual representation. This chapter concludes with a discussion of the challenges associated with FMs and their application in the domain of cybersecurity.

Meta-reinforcement learning (meta-RL) is a promising framework for tackling challenging domains requiring efficient exploration. Existing meta-RL algorithms are characterized by low sample efficiency, and mostly focus on low-dimensional task distributions. In parallel, model-based RL methods have been successful in solving partially observable MDPs, of which meta-RL is a special case. In this work, we leverage this success and propose a new model-based approach to meta-RL, based on elements from existing state-of-the-art model-based and meta-RL methods. We demonstrate the effectiveness of our approach on common meta-RL benchmark domains, attaining greater return with better sample efficiency (up to $15\times$) while requiring very little hyperparameter tuning. In addition, we validate our approach on a slate of more challenging, higher-dimensional domains, taking a step towards real-world generalizing agents.

Concept Bottleneck Model (CBM) is a methods for explaining neural networks. In CBM, concepts which correspond to reasons of outputs are inserted in the last intermediate layer as observed values. It is expected that we can interpret the relationship between the output and concept similar to linear regression. However, this interpretation requires observing all concepts and decreases the generalization performance of neural networks. Partial CBM (PCBM), which uses partially observed concepts, has been devised to resolve these difficulties. Although some numerical experiments suggest that the generalization performance of PCBMs is almost as high as that of the original neural networks, the theoretical behavior of its generalization error has not been yet clarified since PCBM is singular statistical model. In this paper, we reveal the Bayesian generalization error in PCBM with a three-layered and linear architecture. The result indcates that the structure of partially observed concepts decreases the Bayesian generalization error compared with that of CBM (full-observed concepts).

Discriminative methods for learning structured models have enabled wide-spread use of very rich feature representations. However, the computational cost of feature extraction is prohibitive for large-scale or time-sensitive applications, often dominating the cost of inference in the models. Significant efforts have been devoted to sparsity-based model selection to decrease this cost. Such feature selection methods control computation statically and miss the opportunity to finetune feature extraction to each input at run-time. We address the key challenge of learning to control fine-grained feature extraction adaptively, exploiting nonhomogeneity of the data. We propose an architecture that uses a rich feedback loop between extraction and prediction. The run-time control policy is learned using efficient value-function approximation, which adaptively determines the value of information of features at the level of individual variables for each input. We demonstrate significant speedups over state-of-the-art methods on two challenging datasets. For articulated pose estimation in video, we achieve a more accurate state-of-the-art model that is also faster, with similar results on an OCR task.

In this paper, we seek robust policies for uncertain Markov Decision Processes (MDPs). Most robust optimization approaches for these problems have focussed on the computation of maximin policies which maximize the value corresponding to the worst realization of the uncertainty. Recent work has proposed minimax regret as a suitable alternative to the maximin objective for robust optimization. However, existing algorithms for handling minimax regret are restricted to models with uncertainty over rewards only. We provide algorithms that employ sampling to improve across multiple dimensions: (a) Handle uncertainties over both transition and reward models; (b) Dependence of model uncertainties across state, action pairs and decision epochs; (c) Scalability and quality bounds. Finally, to demonstrate the empirical effectiveness of our sampling approaches, we provide comparisons against benchmark algorithms on two domains from literature. We also provide a Sample Average Approximation (SAA) analysis to compute a posteriori error bounds.

This paper gives a spectral algorithm for learning HMM from non-sequential observations. Motivated by several scientific examples, the authors define a sampling model for non-sequential observations that shares some similarities with the generative model of Latent Dirichlet Allocation. Then, resorting to recent spectral techniques for learning LDA, HMM, and mixture models, they prove sample bounds for recovering the parameters of an HMM with continuous output from data sampled according to this model. The last section provides a simple simulation that illustrates the behavior of the algorithm in a synthetic example. Proofs of all results are given in the supplementary material.

Learning dynamic models from observed data has been a central issue in many scientific studies or engineering tasks. The usual setting is that data are collected sequentially from trajectories of some dynamical system operation. In quite a few modern scientific modeling tasks, however, it turns out that reliable sequential data are rather difficult to gather, whereas out-of-order snapshots are much easier to obtain. Examples include the modeling of galaxies, chronic diseases such Alzheimer's, or certain biological processes. Existing methods for learning dynamic model from non-sequence data are mostly based on Expectation-Maximization, which involves non-convex optimization and is thus hard to analyze. Inspired by recent advances in spectral learning methods, we propose to study this problem from a different perspective: moment matching and spectral decomposition. Under that framework, we identify reasonable assumptions on the generative process of non-sequence data, and propose learning algorithms based on the tensor decomposition method [2] to provably recover firstorder Markov models and hidden Markov models. To the best of our knowledge, this is the first formal guarantee on learning from non-sequence data. Preliminary simulation results confirm our theoretical findings.

POMDPs provide a principled framework for planning under uncertainty, but are computationally intractable, due to the "curse of dimensionality" and the "curse of history". This paper presents an online POMDP algorithm that alleviates these difficulties by focusing the search on a set of randomly sampled scenarios.

We investigate the problem of learning the structure of a Markov network from data. It is shown that the structure of such networks can be described in terms of constraints which enables the use of existing solver technology with optimization capabilities to compute optimal networks starting from initial scores computed from the data. To achieve efficient encodings, we develop a novel characterization of Markov network structure using a balancing condition on the separators between cliques forming the network. The resulting translations into propositional satisfiability and its extensions such as maximum satisfiability, satisfiability modulo theories, and answer set programming, enable us to prove optimal certain networks which have been previously found by stochastic search.

The paper provides a simplified perspective on EDML and illustrated the benefits of such a simplified view. First, existing results about EDML are much easier to proce. Second, and most important, this new formulation provides a systematic procedure for deriving new instances of EDML for other models. This is illustrated for Markov networks. The derived EDML is shown to be competitive to existing learning approaches.