Semantic scene understanding allows a robotic agent to reason about problems in complex ways, using information from multiple and varied sensors to make deductions about a particular matter. As a result, this form of intelligent robotics is capable of performing more complex tasks and achieving more precise results than simpler approaches based on single data sources. However, these improved capabilities come at the cost of higher complexity, both computational and in terms of design. Due to the increased design complexity, formal approaches for exploiting semantic understanding become necessary. We present here a probabilistic formulation for integrating semantic knowledge into the process of gas source localization (GSL). The problem of GSL poses many unsolved challenges, and proposed solutions need to contend with the constraining limitations of sensing hardware. By exploiting semantic scene understanding, we can leverage other sources of information, such as vision, to improve the estimation of the source location. We show how our formulation can be applied to pre-existing GSL algorithms and the effect that including semantic data has on the produced estimations of the location of the source.

Causal discovery --the process of identifying cause-and-effect relationships from observational data-- is a pivotal challenge in artificial intelligence (AI) and machine learning. Unveiling causal structures enables robust predictions, facilitates counterfactual reasoning, and enhances decision-making processes in complex systems [1]. Traditional methods for causal discovery often rely on statistical tests for independence and structural equation modeling, which may not scale efficiently with high-dimensional data or effectively capture intricate non-linear relationships [2, 3]. In recent years, machine learning models, particularly deep learning architectures, have achieved remarkable success in predictive tasks. However, these models are typically considered "black boxes" due to their lack of interpretability. This opacity has led to a growing interest in explainable AI (XAI) techniques, with Shapley values emerging as a prominent method for interpreting model predictions [4]. Shapley values, grounded in cooperative game theory, provide a principled approach to attributing the contribution of each feature to the output of a model by quantifying the average marginal contribution of a feature across all possible subsets of features [5]. While Shapley values offer valuable insights into feature importance within a model's predictive framework, the link between feature importance and causal influence is non-trivial.

Self-supervised learning (SSL) aims to find meaningful representations from unlabeled data by encoding semantic similarities through data augmentations. Despite its current popularity, theoretical insights about SSL are still scarce. For example, it is not yet known whether commonly used SSL loss functions can be related to a statistical model, much in the same as OLS, generalized linear models or PCA naturally emerge as maximum likelihood estimates of an underlying generative process. In this short paper, we consider a latent variable statistical model for SSL that exhibits an interesting property: Depending on the informativeness of the data augmentations, the MLE of the model either reduces to PCA, or approaches a simple non-contrastive loss. We analyze the model and also empirically illustrate our findings.

Recent research has highlighted a critical issue known as "robust fairness", where robust accuracy varies significantly across different classes, undermining the reliability of deep neural networks (DNNs). A common approach to address this has been to dynamically reweight classes during training, giving more weight to those with lower empirical robust performance. However, we find there is a divergence of class-wise robust performance between training set and testing set, which limits the effectiveness of these explicit reweighting methods, indicating the need for a principled alternative. In this work, we derive a robust generalization bound for the worst-class robust error within the PAC-Bayesian framework, accounting for unknown data distributions. Our analysis shows that the worst-class robust error is influenced by two main factors: the spectral norm of the empirical robust confusion matrix and the information embedded in the model and training set. While the latter has been extensively studied, we propose a novel regularization technique targeting the spectral norm of the robust confusion matrix to improve worst-class robust accuracy and enhance robust fairness. Deep neural networks, spanning a diverse array of domains and applications, have shown impressive abilities to learn from training data and generalize effectively to new, unseen data. However, recent studies have uncovered a notable weakness in these DNNs - their vulnerability to subtle, often undetectable "adversarial attacks" (Biggio et al., 2013; Szegedy et al., 2013). It has been discovered that even slight perturbations to the input, typically imperceptible to humans, can drastically mislead the networks, resulting in significant prediction errors (Goodfellow et al., 2015; Wu et al., 2020a).

Several scenarios require the optimization of non-convex black-box functions, that are noisy expensive to evaluate functions with unknown analytical expression, whose gradients are hence not accessible. For example, the hyper-parameter tuning problem of machine learning models. Bayesian optimization is a class of methods with state-of-the-art performance delivering a solution to this problem in real scenarios. It uses an iterative process that employs a probabilistic surrogate model, typically a Gaussian process, of the objective function to be optimized computing a posterior predictive distribution of the black-box function. Based on the information given by this posterior predictive distribution, Bayesian optimization includes the computation of an acquisition function that represents, for every input space point, the utility of evaluating that point in the next iteraiton if the objective of the process is to retrieve a global extremum. This paper is a survey of the information theoretical acquisition functions, whose performance typically outperforms the rest of acquisition functions. The main concepts of the field of information theory are also described in detail to make the reader aware of why information theory acquisition functions deliver great results in Bayesian optimization and how can we approximate them when they are intractable. We also cover how information theory acquisition functions can be adapted to complex optimization scenarios such as the multi-objective, constrained, non-myopic, multi-fidelity, parallel and asynchronous settings and provide further lines of research.

It remains difficult to evaluate machine learning classifiers in the absence of a large, labeled dataset. While labeled data can be prohibitively expensive or impossible to obtain, unlabeled data is plentiful. Here, we introduce Semi-Supervised Model Evaluation (SSME), a method that uses both labeled and unlabeled data to evaluate machine learning classifiers. SSME is the first evaluation method to take advantage of the fact that: (i) there are frequently multiple classifiers for the same task, (ii) continuous classifier scores are often available for all classes, and (iii) unlabeled data is often far more plentiful than labeled data. The key idea is to use a semi-supervised mixture model to estimate the joint distribution of ground truth labels and classifier predictions. We can then use this model to estimate any metric that is a function of classifier scores and ground truth labels (e.g., accuracy or expected calibration error). We present experiments in four domains where obtaining large labeled datasets is often impractical: (1) healthcare, (2) content moderation, (3) molecular property prediction, and (4) image annotation. Our results demonstrate that SSME estimates performance more accurately than do competing methods, reducing error by 5.1 relative to using labeled data alone and 2.4 relative to the next best competing method. SSME also improves accuracy when evaluating performance across subsets of the test distribution (e.g., specific demographic subgroups) and when evaluating the performance of language models. Rigorous evaluation is essential to the safe deployment of machine learning classifiers. The standard approach is to measure classifier performance using a large labeled dataset. In practice, however, labeled data is often scarce (Culotta & McCallum, 2005; Dutta & Das, 2023). Exacerbating the challenge of evaluation, the number of off-the-shelf classifiers has increased dramatically through the widespread usage of model hubs. The modern machine learning practitioner thus has a myriad of trained models, but little labeled data with which to evaluate them. In many domains, unlabeled data is much more abundant than labeled data (Bepler et al., 2019; Sagawa et al., 2021; Movva et al., 2024).

The author responses answered my questions as well as points raised by other reviewers, providing additional clarification.] This paper formulates a fully probabilistic model of the vesicle-release dynamics at the sub-cellular biophysical level in the ribbon synapse. The paper then develops a likelihood-free inference method, tests it on a synthetic dataset, and finally infers the parameters of vesicle release in the ribbon synapse from real data. Originality: The paper presents a novel combination of biophysical modeling of ribbon synapse and a likelihood-free inference of the parameters. To my knowledge, the fully stochastic modeling of the vesicle-release dynamics is itself new.

This is an interesting paper on a mechanistic model of the ribbon synapse along with an ABC inference approach. Neither component is particularly novel, but the paper is thorough and compelling. The audience will likely be computationally-savvy experimental neuroscientists and those interested in applications of ABC; the former may be harder to find at NeurIPS, though they do exist. I encourage the authors to make the suggested revisions before the camera ready deadline.

In lines 245-248 the authors discuss a fair comparison between the different methods and mention their effort to keep the total number of coupling layers the same between several methods the same. Can the authors please also comment on the difference in the number of parameters? As the coupling layers in M-Flows don't always act on data of the same dimensionality as regular AF flows, the number of parameters can be different, even with the same number of coupling layers. For the celebA dataset, have you tried to train M-Flows with different n then 512? 4. Can you explain in the main text on a high level why including the SCANDAL loss consistently leads to a larger closure for all methods (lower closure is better). In general, since the supplementary material contains so much more material, it would help the reader if you refer more frequently to the relevant parts of the supplementary material in the main text.

All reviewers agree that the presented technique for simultaneous manifold and density estimation is interesting and novel. However, they also agree that the paper leaves important questions open. While one of the reviewers would like to see a stronger statistical analysis before acceptance, the others believe that the paper is above acceptance threshold and that the community would benefit from its communication. To address the concerns of the reviewers, the camera-ready paper needed to include at least the following results: 1. Include results that investigate if the invertible nature of the normalising flow in the decoder is useful by e.g considering a version of the Me-flow where g is not constrained to be invertible. In the same vein, a comparison with a simple VAE baseline should be included. Investigate how the results on CelebA depend on the latent dimension n.