Fine-grained understanding of objects, attributes, and relationships between objects is crucial for visual-language models (VLMs). To evaluate VLMs' fine-grained understanding, existing benchmarks primarily focus on evaluating VLMs' capability to distinguish between two very similar captions given an image. In this paper, our focus is on evaluating VLMs' capability to distinguish between two very similar images given a caption. To this end, we introduce a new, challenging benchmark termed Visual Minimal-Change Understanding (VisMin), which requires models to predict the correct image-caption match given two images and two captions. Importantly, the image pair (as well as the caption pair) contains minimal changes, i.e., between the two images (as well as between the two captions), only one aspect changes at a time from among the following possible types of changes: object, attribute, count, and spatial relation.

Abstract--Explainability is increasingly seen as an essential feature of rule-based smart environments. While counterfactual explanations, which describe what could have been done differently to achieve a desired outcome, are a powerful tool in eXplainable AI (XAI), no established methods exist for generating them in these rule-based domains. In this paper, we present the first formalization and implementation of counterfactual explanations tailored to this domain. It is implemented as a plugin that extends an existing explanation engine for smart environments. We conducted a user study (N=17) to evaluate our generated counterfactuals against traditional causal explanations. The results show that user preference is highly contextual: causal explanations are favored for their linguistic simplicity and in time-pressured situations, while counterfactuals are preferred for their actionable content, particularly when a user wants to resolve a problem. Our work contributes a practical framework for a new type of explanation in smart environments and provides empirical evidence to guide the choice of when each explanation type is most effective. Smart environments, such as smart homes, offices, and buildings, integrate sensor-enabled devices to support users in decision-making, monitoring, and managing abnormal situations [1], [2]. The rapid adoption of these environments is fueled by advances in the Internet of Things (IoT) and Artificial Intelligence (AI), decreasing device costs, and improved system integration [3]-[5]. Rule-based systems are a prevalent approach for implementing automation in smart environments, by executing predefined rules when certain conditions are met [6], [7].

Within the formal setting of the Lockean thesis, an agent belief set is defined in terms of degrees of confidence and these are described in probabilistic terms. This approach is of established interest, notwithstanding some limitations that make its use troublesome in some contexts, like, for instance, in belief change theory. Precisely, Lockean belief sets are not generally closed under (classical) logical deduction. The aim of the present paper is twofold: on one side we provide two characterizations of those belief sets that are closed under classical logic deduction, and on the other we propose an approach to probabilistic update that allows us for a minimal revision of those beliefs, i.e., a revision obtained by making the fewest possible changes to the existing belief set while still accommodating the new information. In particular, we show how we can deductively close a belief set via a minimal revision.

Fine-grained understanding of objects, attributes, and relationships between objects is crucial for visual-language models (VLMs). To evaluate VLMs' fine-grained understanding, existing benchmarks primarily focus on evaluating VLMs' capability to distinguish between two very similar captions given an image. In this paper, our focus is on evaluating VLMs' capability to distinguish between two very similar images given a caption. To this end, we introduce a new, challenging benchmark termed Visual Minimal-Change Understanding (VisMin), which requires models to predict the correct image-caption match given two images and two captions. Importantly, the image pair (as well as the caption pair) contains minimal changes, i.e., between the two images (as well as between the two captions), only one aspect changes at a time from among the following possible types of changes: object, attribute, count, and spatial relation.

In this paper, we introduce a novel postprocessing algorithm that is both model-agnostic Although rarely discussed, expecting the debiasing method and does not require the sensitive attribute at test to perform a low number of prediction changes is especially time. In addition, our algorithm is explicitly designed interesting in contexts where fairness is enforced while a to enforce minimal changes between biased model is already in production (Krco et al., 2023). In realworld and debiased predictions--a property that, applications, maintaining the integrity and reliability while highly desirable, is rarely prioritized as an of predictive models is crucial, especially when they have explicit objective in fairness literature. Our approach undergone rigorous validation and expert review. For example, leverages a multiplicative factor applied in non-life insurance pricing, experts commonly to the logit value of probability scores produced employ Generalized Additive Models (GAMs) with splines by a black-box classifier. We demonstrate the efficacy or polynomial regression on Generalized Linear Models to of our method through empirical evaluations, ensure that price are justifiable and align with both business comparing its performance against other four debiasing objectives and customer expectations (e.g., avoiding price algorithms on two widely used datasets in increases that could negatively impact customer satisfaction fairness research.

This paper explores an intriguing observation: fine-tuning a large language model (LLM) with responses generated by a LLM often yields better results than using responses generated by humans. We conduct an in-depth investigation to understand why this occurs. Contrary to the common belief that these instances is simply due to the more detailed nature of LLM-generated content, our study identifies another contributing factor: an LLM is inherently more "familiar" with LLM generated responses. This familiarity is evidenced by lower perplexity before fine-tuning. We design a series of experiments to understand the impact of the "familiarity" and our conclusion reveals that this "familiarity" significantly impacts learning performance. Training with LLM-generated responses not only enhances performance but also helps maintain the model's capabilities in other tasks after fine-tuning on a specific task.

In belief revision, agents typically modify their beliefs when they receive some new piece of information that is in conflict with them. The guiding principle behind most belief revision frameworks is that of minimalism, which advocates minimal changes to existing beliefs. However, minimalism may not necessarily capture the nuanced ways in which human agents reevaluate and modify their beliefs. In contrast, the explanatory hypothesis indicates that people are inherently driven to seek explanations for inconsistencies, thereby striving for explanatory coherence rather than minimal changes when revising beliefs. Our contribution in this paper is two-fold. Motivated by the explanatory hypothesis, we first present a novel, yet simple belief revision operator that, given a belief base and an explanation for an explanandum, it revises the belief bases in a manner that preserves the explanandum and is not necessarily minimal. We call this operator explanation-based belief revision. Second, we conduct two human-subject studies to empirically validate our approach and investigate belief revision behavior in real-world scenarios. Our findings support the explanatory hypothesis and provide insights into the strategies people employ when resolving inconsistencies.

As NLP models become more complex, understanding their decisions becomes more crucial. Counterfactuals (CFs), where minimal changes to inputs flip a model's prediction, offer a way to explain these models. While Large Language Models (LLMs) have shown remarkable performance in NLP tasks, their efficacy in generating high-quality CFs remains uncertain. This work fills this gap by investigating how well LLMs generate CFs for two NLU tasks. We conduct a comprehensive comparison of several common LLMs, and evaluate their CFs, assessing both intrinsic metrics, and the impact of these CFs on data augmentation. Moreover, we analyze differences between human and LLM-generated CFs, providing insights for future research directions. Our results show that LLMs generate fluent CFs, but struggle to keep the induced changes minimal. Generating CFs for Sentiment Analysis (SA) is less challenging than NLI where LLMs show weaknesses in generating CFs that flip the original label. This also reflects on the data augmentation performance, where we observe a large gap between augmenting with human and LLMs CFs. Furthermore, we evaluate LLMs' ability to assess CFs in a mislabelled data setting, and show that they have a strong bias towards agreeing with the provided labels. GPT4 is more robust against this bias and its scores correlate well with automatic metrics. Our findings reveal several limitations and point to potential future work directions.

Natural revision seems so natural: it changes beliefs as little as possible to incorporate new information. Yet, some counterexamples show it wrong. It is so conservative that it never fully believes. It only believes in the current conditions. This is right in some cases and wrong in others. Which is which? The answer requires extending natural revision from simple formulae expressing universal truths (something holds) to conditionals expressing conditional truth (something holds in certain conditions). The extension is based on the basic principles natural revision follows, identified as minimal change, indifference and naivety: change beliefs as little as possible; equate the likeliness of scenarios by default; believe all until contradicted. The extension says that natural revision restricts changes to the current conditions. A comparison with an unrestricting revision shows what exactly the current conditions are. It is not what currently considered true if it contradicts the new information. It includes something more and more unlikely until the new information is at least possible.

Decentralized learning has shown great promise for cooperative multi-agent reinforcement learning (MARL). However, non-stationarity remains a significant challenge in fully decentralized learning. In the paper, we tackle the non-stationarity problem in the simplest and fundamental way and propose multi-agent alternate Q-learning (MA2QL), where agents take turns updating their Q-functions by Q-learning. MA2QL is a minimalist approach to fully decentralized cooperative MARL but is theoretically grounded. We prove that when each agent guarantees $\varepsilon$-convergence at each turn, their joint policy converges to a Nash equilibrium. In practice, MA2QL only requires minimal changes to independent Q-learning (IQL). We empirically evaluate MA2QL on a variety of cooperative multi-agent tasks. Results show MA2QL consistently outperforms IQL, which verifies the effectiveness of MA2QL, despite such minimal changes.