Concept-based explainable artificial intelligence (C-XAI) can let people see which representations an AI model has learned. This is particularly important when high-level semantic information (e.g., actions and relations) is used to make decisions about abstract categories (e.g., danger). In such tasks, AI models need to generalise beyond situation-specific details, and this ability can be reflected in C-XAI outputs that randomise over irrelevant features. However, it is unclear whether people appreciate such generalisation and can distinguish it from other, less desirable forms of imprecision in C-XAI outputs. Therefore, the present study investigated how the generality and relevance of C-XAI outputs affect people's evaluation of AI. In an experimental railway safety evaluation scenario, participants rated the performance of a simulated AI that classified traffic scenes involving people as dangerous or not. These classification decisions were explained via concepts in the form of similar image snippets. The latter differed in their match with the classified image, either regarding a highly relevant feature (i.e., people's relation to tracks) or a less relevant feature (i.e., people's action). Contrary to the hypotheses, concepts that generalised over less relevant features were rated lower than concepts that matched the classified image precisely. Moreover, their ratings were no better than those for systematic misrepresentations of the less relevant feature. Conversely, participants were highly sensitive to imprecisions in relevant features. These findings cast doubts on the assumption that people can easily infer from C-XAI outputs whether AI models have gained a deeper understanding of complex situations.

Recent research has attempted to associate preference optimization (PO) performance with the underlying preference datasets. In this work, our observation is that the differences between the preferred response $y^+$ and dispreferred response $y^-$ influence what LLMs can learn, which may not match the desirable differences to learn. Therefore, we use distance and reward margin to quantify these differences, and combine them to get Distance Calibrated Reward Margin (DCRM), a metric that measures the quality of a response pair for PO. Intuitively, DCRM encourages minimal noisy differences and maximal desired differences. With this, we study 3 types of commonly used preference datasets, classified along two axes: the source of the responses and the preference labeling function. We establish a general correlation between higher DCRM of the training set and better learning outcome. Inspired by this, we propose a best-of-$N^2$ pairing method that selects response pairs with the highest DCRM. Empirically, in various settings, our method produces training datasets that can further improve models' performance on AlpacaEval, MT-Bench, and Arena-Hard over the existing training sets.

Tree ensembles such as Random Forests have achieved impressive empirical success across a wide variety of applications.

Deep learning has achieved remarkable success across domains such as vision, language, and science. A widely believed explanation for this success is representation learning -- also called feature learning -- the empirically observed ability of deep models to automatically extract task-relevant features from raw data, without manual engineering, to support downstream prediction [1]. This ability is generally attributed to two fundamental ingredients of deep models: (i) their compositional architecture and (ii) the use of optimization. The compositionality of the architecture endows the model with the ability to form intermediate representations of the data via composition of simple transformations. These representations are not manually defined but are learned from data by optimizing a loss function designed to minimize prediction error. However, despite the empirical success of this paradigm, our theoretical understanding of how and why such representations emerge remains fundamentally limited. In particular, it remains unclear how the interplay between compositional structure and optimization gives rise to task-aligned features -- and under what conditions this mechanism succeeds or fails. To address this gap, we study a stylized compositional model that preserves these two core ingredients of feature learning -- while remaining simple enough to enable analysis of how features are learnt during training.

"NIPS 2013 Neural Information Processing Systems December 5 - 10, Lake Tahoe, Nevada, USA",,, "Paper ID:","1051" "Title:","Minimax Theory for High-dimensional Gaussian Mixtures with Sparse Mean Separation" Reviews First provide a summary of the paper, and then address the following criteria: Quality, clarity, originality and significance. The paper studies the problem of identifying Gaussians in a mixture in high dimensions when the separation between the Gaussians is small. The assumption is that the Gaussians are separated along few dimensions and hence by identifying these dimensions, that is, feature selection, the curse of dimensionality can be bitten and the Gaussians can be found. Clustering in high dimension is an open problem that well deserve a study. The theoretical approach taken by the authors is good step in the path towards better understanding the problem.

First provide a summary of the paper, and then address the following criteria: Quality, clarity, originality and significance. This paper studies the multi-armed bandit problem where they have a set of relevant features; and the expected reward of an action is a Lipschitz continuous of relevant features. This is also a feature selection problem where you have a set of features but only r of them are relevant (the target function only depends on r of these features): here each arm has only one relevant feature, meaning the function representing the arm payoff depending on only one feature and we do not know which one. They propose an algorithm and get the bound for such adaptive case; but their regret is higher than what you would get if someone tells you the relevant type. Q2: Please summarize your review in 1-2 sentences This paper makes a small step towards understanding the problem of having a subset of features being relevant for a given arm which itself is certainly an interesting problem: they study the bandit problem only for one relevant feature per arm and did not give the optimal rate. Potentially, they could go with all arbitrary number of relevant features and figure out the optimal regret.

Latent representations are critical for the performance and robustness of machine learning models, as they encode the essential features of data in a compact and informative manner. However, in vision tasks, these representations are often affected by noisy or irrelevant features, which can degrade the model's performance and generalization capabilities. This paper presents a novel approach for enhancing latent representations using unsupervised Dynamic Feature Selection (DFS). For each instance, the proposed method identifies and removes misleading or redundant information in images, ensuring that only the most relevant features contribute to the latent space. By leveraging an unsupervised framework, our approach avoids reliance on labeled data, making it broadly applicable across various domains and datasets. Experiments conducted on image datasets demonstrate that models equipped with unsupervised DFS achieve significant improvements in generalization performance across various tasks, including clustering and image generation, while incurring a minimal increase in the computational cost.

Neural Information Processing Systems http://nips.cc/