Optimizing static risk-averse objectives in Markov decision processes is difficult because they do not admit standard dynamic programming equations common in Reinforcement Learning (RL) algorithms. Dynamic programming decompositions that augment the state space with discrete risk levels have recently gained popularity in the RL community. Prior work has shown that these decompositions are optimal when the risk level is discretized sufficiently. However, we show that these popular decompositions for Conditional-Value-at-Risk (CVaR) and Entropic-Value-at-Risk (EVaR) are inherently suboptimal regardless of the discretization level. In particular, we show that a saddle point property assumed to hold in prior literature may be violated. However, a decomposition does hold for Value-at-Risk and our proof demonstrates how this risk measure differs from CVaR and EVaR. Our findings are significant because risk-averse algorithms are used in high-stake environments, making their correctness much more critical.
相關內容
This manuscript presents a novel method for discovering effective connectivity between specified pairs of nodes in a high-dimensional network of time series. To accurately perform Granger causality analysis from the first node to the second node, it is essential to eliminate the influence of all other nodes within the network. The approach proposed is to create a low-dimensional representation of all other nodes in the network using frequency-domain-based dynamic principal component analysis (spectral DPCA). The resulting scores are subsequently removed from the first and second nodes of interest, thus eliminating the confounding effect of other nodes within the high-dimensional network. To conduct hypothesis testing on Granger causality, we propose a permutation-based causality test. This test enhances the accuracy of our findings when the error structures are non-Gaussian. The approach has been validated in extensive simulation studies, which demonstrate the efficacy of the methodology as a tool for causality analysis in complex time series networks. The proposed methodology has also been demonstrated to be both expedient and viable on real datasets, with particular success observed on multichannel EEG networks.
Combined electric power system and High-Altitude Electromagnetic Pulse (HEMP) models are being developed to determine the effect of a HEMP on the US power grid. The work relies primarily on deterministic methods; however, it is computationally untenable to evaluate the E1 HEMP response of large numbers of grid components distributed across a large interconnection. Further, the deterministic assessment of these components' failures are largely unachievable. E1 HEMP laboratory testing of the components is accomplished, but is expensive, leaving few data points to construct failure models of grid components exposed to E1 HEMP. The use of Bayesian priors, developed using the subject matter expertise, combined with the minimal test data in a Bayesian inference process, provides the basis for the development of more robust and cost-effective statistical component failure models. These can be used with minimal computational burden in a simulation environment such as sampling of Cumulative Distribution Functions (CDFs).
Copyright infringement may occur when a generative model produces samples substantially similar to some copyrighted data that it had access to during the training phase. The notion of access usually refers to including copyrighted samples directly in the training dataset, which one may inspect to identify an infringement. We argue that such visual auditing largely overlooks a concealed copyright infringement, where one constructs a disguise that looks drastically different from the copyrighted sample yet still induces the effect of training Latent Diffusion Models on it. Such disguises only require indirect access to the copyrighted material and cannot be visually distinguished, thus easily circumventing the current auditing tools. In this paper, we provide a better understanding of such disguised copyright infringement by uncovering the disguises generation algorithm, the revelation of the disguises, and importantly, how to detect them to augment the existing toolbox. Additionally, we introduce a broader notion of acknowledgment for comprehending such indirect access. Our code is available at //github.com/watml/disguised_copyright_infringement.
An important aspect in developing language models that interact with humans is aligning their behavior to be useful and unharmful for their human users. This is usually achieved by tuning the model in a way that enhances desired behaviors and inhibits undesired ones, a process referred to as alignment. In this paper, we propose a theoretical approach called Behavior Expectation Bounds (BEB) which allows us to formally investigate several inherent characteristics and limitations of alignment in large language models. Importantly, we prove that within the limits of this framework, for any behavior that has a finite probability of being exhibited by the model, there exist prompts that can trigger the model into outputting this behavior, with probability that increases with the length of the prompt. This implies that any alignment process that attenuates an undesired behavior but does not remove it altogether, is not safe against adversarial prompting attacks. Furthermore, our framework hints at the mechanism by which leading alignment approaches such as reinforcement learning from human feedback make the LLM prone to being prompted into the undesired behaviors. This theoretical result is being experimentally demonstrated in large scale by the so called contemporary "chatGPT jailbreaks", where adversarial users trick the LLM into breaking its alignment guardrails by triggering it into acting as a malicious persona. Our results expose fundamental limitations in alignment of LLMs and bring to the forefront the need to devise reliable mechanisms for ensuring AI safety.
In the quantum computation verification problem, a quantum server wants to convince a client that the output of evaluating a quantum circuit $C$ is some result that it claims. This problem is considered very important both theoretically and practically in quantum computation [arXiv:1709.06984], [arXiv:1704.04487], [arXiv:1209.0449]. The client is considered to be limited in computational power, and one desirable property is that the client can be completely classical, which leads to the classical verification of quantum computation (CVQC) problem. In terms of the total time complexity, the fastest single-server CVQC protocol so far has complexity $O(poly(\kappa)|C|^3)$ where $|C|$ is the size of the circuit to be verified and $\kappa$ is the security parameter, given by Mahadev [arXiv:1804.01082]. In this work, by developing new techniques, we give a new CVQC protocol with complexity $O(poly(\kappa)|C|)$, which is significantly faster than existing protocols. Our protocol is secure in the quantum random oracle model [arXiv:1008.0931] assuming the existence of noisy trapdoor claw-free functions [arXiv:1804.00640], which are both extensively used assumptions in quantum cryptography. Along the way, we also give a new classical channel remote state preparation protocol for states in $\{|+_\theta\rangle=\frac{1}{\sqrt{2}}(|0\rangle+e^{i\theta\pi/4}|1\rangle):\theta\in \{0,1\cdots 7\}\}$, another basic primitive in quantum cryptography. Our protocol allows for parallel verifiable preparation of $L$ independently random states in this form (up to a constant overall error and a possibly unbounded server-side simulator), and runs in only $O(poly(\kappa)L)$ time and constant rounds; for comparison, existing works (even for possibly simpler state families) all require very large or unestimated time and round complexities [arXiv:1904.06320][arXiv:1904.06303][arXiv:2201.13445][arXiv:2201.13430].
Quantum computing represents a revolutionary computational paradigm with the potential to address challenges beyond classical computers' capabilities. The development of robust quantum software is indispensable to unlock the full potential of quantum computing. Like classical software, quantum software is expected to be complex and extensive, needing the establishment of a specialized field known as Quantum Software Engineering. Recognizing the regional focus on Latin America within this special issue, we have boarded on an in-depth inquiry encompassing a systematic mapping study of existing literature and a comprehensive survey of experts in the field. This rigorous research effort aims to illuminate the current landscape of Quantum Software Engineering initiatives undertaken by universities, research institutes, and companies across Latin America. This exhaustive study aims to provide information on the progress, challenges, and opportunities in Quantum Software Engineering in the Latin American context. By promoting a more in-depth understanding of cutting-edge developments in this burgeoning field, our research aims to serve as a potential stimulus to initiate pioneering initiatives and encourage collaborative efforts among Latin American researchers.
As artificial intelligence (AI) models continue to scale up, they are becoming more capable and integrated into various forms of decision-making systems. For models involved in moral decision-making, also known as artificial moral agents (AMA), interpretability provides a way to trust and understand the agent's internal reasoning mechanisms for effective use and error correction. In this paper, we provide an overview of this rapidly-evolving sub-field of AI interpretability, introduce the concept of the Minimum Level of Interpretability (MLI) and recommend an MLI for various types of agents, to aid their safe deployment in real-world settings.
The existence of representative datasets is a prerequisite of many successful artificial intelligence and machine learning models. However, the subsequent application of these models often involves scenarios that are inadequately represented in the data used for training. The reasons for this are manifold and range from time and cost constraints to ethical considerations. As a consequence, the reliable use of these models, especially in safety-critical applications, is a huge challenge. Leveraging additional, already existing sources of knowledge is key to overcome the limitations of purely data-driven approaches, and eventually to increase the generalization capability of these models. Furthermore, predictions that conform with knowledge are crucial for making trustworthy and safe decisions even in underrepresented scenarios. This work provides an overview of existing techniques and methods in the literature that combine data-based models with existing knowledge. The identified approaches are structured according to the categories integration, extraction and conformity. Special attention is given to applications in the field of autonomous driving.
Object detection typically assumes that training and test data are drawn from an identical distribution, which, however, does not always hold in practice. Such a distribution mismatch will lead to a significant performance drop. In this work, we aim to improve the cross-domain robustness of object detection. We tackle the domain shift on two levels: 1) the image-level shift, such as image style, illumination, etc, and 2) the instance-level shift, such as object appearance, size, etc. We build our approach based on the recent state-of-the-art Faster R-CNN model, and design two domain adaptation components, on image level and instance level, to reduce the domain discrepancy. The two domain adaptation components are based on H-divergence theory, and are implemented by learning a domain classifier in adversarial training manner. The domain classifiers on different levels are further reinforced with a consistency regularization to learn a domain-invariant region proposal network (RPN) in the Faster R-CNN model. We evaluate our newly proposed approach using multiple datasets including Cityscapes, KITTI, SIM10K, etc. The results demonstrate the effectiveness of our proposed approach for robust object detection in various domain shift scenarios.
Detecting carried objects is one of the requirements for developing systems to reason about activities involving people and objects. We present an approach to detect carried objects from a single video frame with a novel method that incorporates features from multiple scales. Initially, a foreground mask in a video frame is segmented into multi-scale superpixels. Then the human-like regions in the segmented area are identified by matching a set of extracted features from superpixels against learned features in a codebook. A carried object probability map is generated using the complement of the matching probabilities of superpixels to human-like regions and background information. A group of superpixels with high carried object probability and strong edge support is then merged to obtain the shape of the carried object. We applied our method to two challenging datasets, and results show that our method is competitive with or better than the state-of-the-art.
小貼士
登錄享
注冊地址： 北京市海淀區羊坊店路18號2幢3層301-191