Off-policy evaluation (OPE) is a critical challenge in robust decision-making that seeks to assess the performance of a new policy using data collected under a different policy. However, the existing OPE methodologies suffer from several limitations arising from statistical uncertainty as well as causal considerations. In this thesis, we address these limitations by presenting three different works. Firstly, we consider the problem of high variance in the importance-sampling-based OPE estimators. We introduce the Marginal Ratio (MR) estimator, a novel OPE method that reduces variance by focusing on the marginal distribution of outcomes rather than direct policy shifts, improving robustness in contextual bandits. Next, we propose Conformal Off-Policy Prediction (COPP), a principled approach for uncertainty quantification in OPE that provides finite-sample predictive intervals, ensuring robust decision-making in risk-sensitive applications. Finally, we address causal unidentifiability in off-policy decision-making by developing novel bounds for sequential decision settings, which remain valid under arbitrary unmeasured confounding. We apply these bounds to assess the reliability of digital twin models, introducing a falsification framework to identify scenarios where model predictions diverge from real-world behaviour. Our contributions provide new insights into robust decision-making under uncertainty and establish principled methods for evaluating policies in both static and dynamic settings.
相關內容
We study the tasks of collective coin flipping and leader election in the full-information model. We prove new lower bounds for coin flipping protocols, implying lower bounds for leader election protocols. We show that any $k$-round coin flipping protocol, where each of $\ell$ players sends 1 bit per round, can be biased by $O(\ell/\log^{(k)}(\ell))$ bad players. For all $k>1$ this strengthens previous lower bounds [RSZ, SICOMP 2002], which ruled out protocols resilient to adversaries controlling $O(\ell/\log^{(2k-1)}(\ell))$ players. Consequently, we establish that any protocol tolerating a linear fraction of corrupt players, with only 1 bit per round, must run for at least $\log^*\ell-O(1)$ rounds, improving on the prior best lower bound of $\frac12 \log^*\ell-\log^*\log^*\ell$. This lower bound matches the number of rounds, $\log^*\ell$, taken by the current best coin flipping protocols from [RZ, JCSS 2001], [F, FOCS 1999] that can handle a linear sized coalition of bad players, but with players sending unlimited bits per round. We also derive lower bounds for protocols allowing multi-bit messages per round. Our results show that the protocols from [RZ, JCSS 2001], [F, FOCS 1999] that handle a linear number of corrupt players are almost optimal in terms of round complexity and communication per player in a round. A key technical ingredient in proving our lower bounds is a new result regarding biasing most functions from a family of functions using a common set of bad players and a small specialized set of bad players specific to each function that is biased. We give improved constant-round coin flipping protocols in the setting that each player can send 1 bit per round. For two rounds, our protocol can handle $O(\ell/(\log\ell)(\log\log\ell)^2)$ sized coalition of bad players; better than the best one-round protocol by [AL, Combinatorica 1993] in this setting.
We investigate the relationship between the team-optimal solution and the Nash equilibrium (NE) to assess the impact of strategy deviation on team performance. As a working use case, we focus on a class of flow assignment problems in which each source node acts as a cooperating decision maker (DM) within a team that minimizes the team cost based on the team-optimal strategy. In practice, some selfish DMs may prioritize their own marginal cost and deviate from NE strategies, thus potentially degrading the overall performance. To quantify this deviation, we explore the deviation bound between the team-optimal solution and the NE in two specific scenarios: (i) when the team-optimal solution is unique and (ii) when multiple solutions do exist. This helps DMs analyze the factors influencing the deviation and adopting the NE strategy within a tolerable range. Furthermore, in the special case of a potential game model, we establish the consistency between the team-optimal solution and the NE. Once the consistency condition is satisfied, the strategy deviation does not alter the total cost, and DMs do not face a strategic trade-off. Finally, we validate our theoretical analysis through some simulation studies.
Semantic communication has emerged as a promising technology for enhancing communication efficiency. However, most existing research emphasizes single-task reconstruction, neglecting model adaptability and generalization across multi-task systems. In this paper, we propose a novel generative semantic communication system that supports both image reconstruction and segmentation tasks. Our approach builds upon semantic knowledge bases (KBs) at both the transmitter and receiver, with each semantic KB comprising a source KB and a task KB. The source KB at the transmitter leverages a hierarchical Swin-Transformer, a generative AI scheme, to extract multi-level features from the input image. Concurrently, the counterpart source KB at the receiver utilizes hierarchical residual blocks to generate task-specific knowledge. Furthermore, the task KBs adopt a semantic similarity model to map different task requirements into pre-defined task instructions, thereby facilitating the feature selection of the source KBs. Additionally, we develop a unified residual block-based joint source and channel (JSCC) encoder and two task-specific JSCC decoders to achieve the two image tasks. In particular, a generative diffusion model is adopted to construct the JSCC decoder for the image reconstruction task. Experimental results show that our multi-task generative semantic communication system outperforms previous single-task communication systems in terms of peak signal-to-noise ratio and segmentation accuracy.
DeepSeek-V3, a recently emerging Large Language Model (LLM), demonstrates outstanding performance in general scene understanding, question-answering (QA), and text generation tasks, owing to its efficient training paradigm and strong reasoning capabilities. In this study, we investigate the dialogue capabilities of DeepSeek-V3 in robotic surgery scenarios, focusing on tasks such as Single Phrase QA, Visual QA, and Detailed Description. The Single Phrase QA tasks further include sub-tasks such as surgical instrument recognition, action understanding, and spatial position analysis. We conduct extensive evaluations using publicly available datasets, including EndoVis18 and CholecT50, along with their corresponding dialogue data. Our comprehensive evaluation results indicate that, when provided with specific prompts, DeepSeek-V3 performs well in surgical instrument and tissue recognition tasks However, DeepSeek-V3 exhibits significant limitations in spatial position analysis and struggles to understand surgical actions accurately. Additionally, our findings reveal that, under general prompts, DeepSeek-V3 lacks the ability to effectively analyze global surgical concepts and fails to provide detailed insights into surgical scenarios. Based on our observations, we argue that the DeepSeek-V3 is not ready for vision-language tasks in surgical contexts without fine-tuning on surgery-specific datasets.
Many extreme-scale applications require the movement of large quantities of data to, from, and among leadership computing facilities, as well as other scientific facilities and the home institutions of facility users. These applications, particularly when leadership computing facilities are involved, can touch upon edge cases (e.g., terabyte files) that had not been a focus of previous Globus optimization work, which had emphasized rather the movement of many smaller (megabyte to gigabyte) files. We report here on how automated client-driven chunking can be used to accelerate both the movement of large files and the integrity checking operations that have proven to be essential for large data transfers. We present detailed performance studies that provide insights into the benefits of these modifications in a range of file transfer scenarios.
Reinforcement Learning (RL) has shown excellent performance in solving decision-making and control problems of autonomous driving, which is increasingly applied in diverse driving scenarios. However, driving is a multi-attribute problem, leading to challenges in achieving multi-objective compatibility for current RL methods, especially in both policy execution and policy iteration. On the one hand, the common action space structure with single action type limits driving flexibility or results in large behavior fluctuations during policy execution. On the other hand, the multi-attribute weighted single reward function result in the agent's disproportionate attention to certain objectives during policy iterations. To this end, we propose a Multi-objective Ensemble-Critic reinforcement learning method with Hybrid Parametrized Action for multi-objective compatible autonomous driving. Specifically, a parameterized action space is constructed to generate hybrid driving actions, combining both abstract guidance and concrete control commands. A multi-objective critics architecture is constructed considering multiple attribute rewards, to ensure simultaneously focusing on different driving objectives. Additionally, uncertainty-based exploration strategy is introduced to help the agent faster approach viable driving policy. The experimental results in both the simulated traffic environment and the HighD dataset demonstrate that our method can achieve multi-objective compatible autonomous driving in terms of driving efficiency, action consistency, and safety. It enhances the general performance of the driving while significantly increasing training efficiency.
This study presents a Bayesian spatial voting analysis of the Colombian Senate during the 2006-2010 legislative period, leveraging a newly constructed roll-call dataset comprising 147 senators and 136 plenary votes. We estimate legislators' ideal points under two alternative geometric frameworks: A traditional Euclidean model and a circular model that embeds preferences on the unit circle. Both models are implemented using Markov Chain Monte Carlo methods, with the circular specification capturing geodesic distances and von Mises-distributed latent traits. The results reveal a latent structure in voting behavior best characterized not by a conventional left-right ideological continuum but by an opposition-non-opposition alignment. Using Bayesian logistic regression, we further investigate the association between senators' ideal points and their involvement in the para-politics scandal. Findings indicate a significant and robust relationship between political alignment and para-politics implication, suggesting that extralegal influence was systematically related to senators' legislative behavior during this period.
Cybersecurity demands rigorous and scalable techniques to ensure system correctness, robustness, and resilience against evolving threats. Automated reasoning, encompassing formal logic, theorem proving, model checking, and symbolic analysis, provides a foundational framework for verifying security properties across diverse domains such as access control, protocol design, vulnerability detection, and adversarial modeling. This survey presents a comprehensive overview of the role of automated reasoning in cybersecurity, analyzing how logical systems, including temporal, deontic, and epistemic logics are employed to formalize and verify security guarantees. We examine SOTA tools and frameworks, explore integrations with AI for neural-symbolic reasoning, and highlight critical research gaps, particularly in scalability, compositionality, and multi-layered security modeling. The paper concludes with a set of well-grounded future research directions, aiming to foster the development of secure systems through formal, automated, and explainable reasoning techniques.
Few-shot Knowledge Graph (KG) completion is a focus of current research, where each task aims at querying unseen facts of a relation given its few-shot reference entity pairs. Recent attempts solve this problem by learning static representations of entities and references, ignoring their dynamic properties, i.e., entities may exhibit diverse roles within task relations, and references may make different contributions to queries. This work proposes an adaptive attentional network for few-shot KG completion by learning adaptive entity and reference representations. Specifically, entities are modeled by an adaptive neighbor encoder to discern their task-oriented roles, while references are modeled by an adaptive query-aware aggregator to differentiate their contributions. Through the attention mechanism, both entities and references can capture their fine-grained semantic meanings, and thus render more expressive representations. This will be more predictive for knowledge acquisition in the few-shot scenario. Evaluation in link prediction on two public datasets shows that our approach achieves new state-of-the-art results with different few-shot sizes.
Multi-relation Question Answering is a challenging task, due to the requirement of elaborated analysis on questions and reasoning over multiple fact triples in knowledge base. In this paper, we present a novel model called Interpretable Reasoning Network that employs an interpretable, hop-by-hop reasoning process for question answering. The model dynamically decides which part of an input question should be analyzed at each hop; predicts a relation that corresponds to the current parsed results; utilizes the predicted relation to update the question representation and the state of the reasoning process; and then drives the next-hop reasoning. Experiments show that our model yields state-of-the-art results on two datasets. More interestingly, the model can offer traceable and observable intermediate predictions for reasoning analysis and failure diagnosis, thereby allowing manual manipulation in predicting the final answer.
注冊地址： 北京市海淀區羊坊店路18號2幢3層301-191