Ensuring both transparency and safety is critical when deploying Deep Neural Networks (DNNs) in high-risk applications, such as medicine. The field of explainable AI (XAI) has proposed various methods to comprehend the decision-making processes of opaque DNNs. However, only few XAI methods are suitable of ensuring safety in practice as they heavily rely on repeated labor-intensive and possibly biased human assessment. In this work, we present a novel post-hoc concept-based XAI framework that conveys besides instance-wise (local) also class-wise (global) decision-making strategies via prototypes. What sets our approach apart is the combination of local and global strategies, enabling a clearer understanding of the (dis-)similarities in model decisions compared to the expected (prototypical) concept use, ultimately reducing the dependence on human long-term assessment. Quantifying the deviation from prototypical behavior not only allows to associate predictions with specific model sub-strategies but also to detect outlier behavior. As such, our approach constitutes an intuitive and explainable tool for model validation. We demonstrate the effectiveness of our approach in identifying out-of-distribution samples, spurious model behavior and data quality issues across three datasets (ImageNet, CUB-200, and CIFAR-10) utilizing VGG, ResNet, and EfficientNet architectures. Code is available on https://github.com/maxdreyer/pcx.

In this paper, we formulate the outfit completion problem as a set retrieval task and propose a novel framework for solving this problem. The proposal includes a conditional set transformation architecture with deep neural networks and a compatibility-based regularization method. The proposed method utilizes a map with permutation-invariant for the input set and permutation-equivariant for the condition set. This allows retrieving a set that is compatible with the input set while reflecting the properties of the condition set. In addition, since this structure outputs the element of the output set in a single inference, it can achieve a scalable inference speed with respect to the cardinality of the output set. Experimental results on real data reveal that the proposed method outperforms existing approaches in terms of accuracy of the outfit completion task, condition satisfaction, and compatibility of completion results.

The development of causal prediction models is challenged by the fact that the outcome is only observable for the applied (factual) intervention and not for its alternatives (the so-called counterfactuals); in medicine we only know patients' survival for the administered drug and not for other therapeutic options. Machine learning approaches for counterfactual reasoning have to deal with both unobserved outcomes and distributional differences due to non-random treatment administration. Unsupervised domain adaptation (UDA) addresses similar issues; one has to deal with unobserved outcomes -- the labels of the target domain -- and distributional differences between source and target domain. We propose Adversarial Distribution Balancing for Counterfactual Reasoning (ADBCR), which directly uses potential outcome estimates of the counterfactuals to remove spurious causal relations. We show that ADBCR outcompetes state-of-the-art methods on three benchmark datasets, and demonstrate that ADBCR's performance can be further improved if unlabeled validation data are included in the training procedure to better adapt the model to the validation domain.

--Knowledge distillation (KD) can enable collaborative learning among distributed clients that have different model architectures and do not share their local data and model parameters with others. Each client updates its local model using the average model output/feature of all client models as the target, known as federated KD. However, existing federated KD methods often do not perform well when clients' local models are trained with heterogeneous local datasets. In this paper, we propose Federated knowledge distillation enabled by Adversarial Learning ( FedAL) to address the data heterogeneity among clients. First, to alleviate the local model output divergence across clients caused by data heterogeneity, the server acts as a discriminator to guide clients' local model training to achieve consensus model outputs among clients through a min-max game between clients and the discriminator . Moreover, catastrophic forgetting may happen during the clients' local training and global knowledge transfer due to clients' heterogeneous local data. T owards this challenge, we design the less-forgetting regularization for both local training and global knowledge transfer to guarantee clients' ability to transfer/learn knowledge to/from others. Experimental results show that FedAL and its variants achieve higher accuracy than other federated KD baselines. To address this problem, collaborative learning among multiple clients can be useful for producing models with better accuracy. However, there are several challenges. First, clients have their own local datasets and they may not be willing to share their raw data with others due to privacy concerns [1]. Second, clients on the edge of wireless networks often have different computation and memory resources, resulting in clients with heterogeneous models that have different architectures and parameters. Clients may not want to reveal their model architectures to other clients to further prevent privacy leakage [2], [3]. We refer to a client's model with unknown architecture to other clients as a black-box model. Pengchao Han is with the School of Information Engineering, Guangdong University of Technology, Guangzhou 510006, China. Han's contribution to this work was made when she was a Postdoc research associate at The Chinese University of Hong Kong, Shenzhen, China.

Balancing the trade-off between accuracy and robustness is a long-standing challenge in time series forecasting. While most of existing robust algorithms have achieved certain suboptimal performance on clean data, sustaining the same performance level in the presence of data perturbations remains extremely hard. In this paper, we study a wide array of perturbation scenarios and propose novel defense mechanisms against adversarial attacks using real-world telecom data. We compare our strategy against two existing adversarial training algorithms under a range of maximal allowed perturbations, defined using $\ell_{\infty}$-norm, $\in [0.1,0.4]$. Our findings reveal that our hybrid strategy, which is composed of a classifier to detect adversarial examples, a denoiser to eliminate noise from the perturbed data samples, and a standard forecaster, achieves the best performance on both clean and perturbed data. Our optimal model can retain up to $92.02\%$ the performance of the original forecasting model in terms of Mean Squared Error (MSE) on clean data, while being more robust than the standard adversarially trained models on perturbed data. Its MSE is 2.71$\times$ and 2.51$\times$ lower than those of comparing methods on normal and perturbed data, respectively. In addition, the components of our models can be trained in parallel, resulting in better computational efficiency. Our results indicate that we can optimally balance the trade-off between the performance and robustness of forecasting models by improving the classifier and denoiser, even in the presence of sophisticated and destructive poisoning attacks.

Let it Care PariSanté Campus, Paris, France PariSanté Campus, Paris, France archibald.fraikin@inria.fr Multivariate time series present challenges to standard machine learning techniques, as they are often unlabeled, high dimensional, noisy, and contain missing data. To address this, we propose T-Rep, a self-supervised method to learn time series representations at a timestep granularity. T-Rep learns vector embeddings of time alongside its feature extractor, to extract temporal features such as trend, periodicity, or distribution shifts from the signal. These time-embeddings are leveraged in pretext tasks, to incorporate smooth and fine-grained temporal dependencies in the representations, as well as reinforce robustness to missing data. We evaluate T-Rep on downstream classification, forecasting, and anomaly detection tasks. It is compared to existing self-supervised algorithms for time series, which it outperforms in all three tasks. We test T-Rep in missing data regimes, where it proves more resilient than its counterparts. Finally, we provide latent space visualisation experiments, highlighting the interpretability of the learned representations. Multivariate time series have become ubiquitous in domains such as medicine, climate science, or finance. Unfortunately, they are high-dimensional and complex objects with little data being labeled (Yang & Wu, 2006), as it is an expensive and time-consuming process. Leveraging unlabeled data to build unsupervised representations of multivariate time series has thus become a challenge of great interest, as these embeddings can significantly improve performance in tasks like forecasting, classification, or anomaly detection (Deldari et al., 2021; Su et al., 2019). This has motivated the development of self-supervised learning (SSL) models for time series, first focusing on constructing instance-level representations for classification and clustering (Tonekaboni et al., 2021; Franceschi et al., 2019; Wu et al., 2018). More fine-grained representations were then developed to model time series at the timestep-level (Yue et al., 2022), which is key in domains such as healthcare or sensor systems. With fine-grained embeddings, one can capture subtle changes, periodic patterns, and irregularities that are essential for anomaly detection (Keogh et al., 2006) as well as understanding and forecasting disease progression.

Algorithmic fairness is of utmost societal importance, yet the current trend in large-scale machine learning models requires training with massive datasets that are frequently biased. In this context, pre-processing methods that focus on modeling and correcting bias in the data emerge as valuable approaches. In this paper, we propose FairShap, a novel instance-level data re-weighting method for fair algorithmic decision-making through data valuation by means of Shapley Values. FairShap is model-agnostic and easily interpretable, as it measures the contribution of each training data point to a predefined fairness metric. We empirically validate FairShap on several state-of-the-art datasets of different nature, with a variety of training scenarios and models and show how it yields fairer models with similar levels of accuracy than the baselines. We illustrate FairShap's interpretability by means of histograms and latent space visualizations. Moreover, we perform a utility-fairness study, and ablation and runtime experiments to illustrate the impact of the size of the reference dataset and FairShap's computational cost depending on the size of the dataset and the number of features. We believe that FairShap represents a promising direction in interpretable and model-agnostic approaches to algorithmic fairness that yield competitive accuracy even when only biased datasets are available.

Exemplar-free class-incremental learning is very challenging due to the negative effect of catastrophic forgetting. A balance between stability and plasticity of the incremental process is needed in order to obtain good accuracy for past as well as new classes. Existing exemplar-free class-incremental methods focus either on successive fine tuning of the model, thus favoring plasticity, or on using a feature extractor fixed after the initial incremental state, thus favoring stability. We introduce a method which combines a fixed feature extractor and a pseudo-features generator to improve the stability-plasticity balance. The generator uses a simple yet effective geometric translation of new class features to create representations of past classes, made of pseudo-features. The translation of features only requires the storage of the centroid representations of past classes to produce their pseudo-features. Actual features of new classes and pseudo-features of past classes are fed into a linear classifier which is trained incrementally to discriminate between all classes. The incremental process is much faster with the proposed method compared to mainstream ones which update the entire deep model. Experiments are performed with three challenging datasets, and different incremental settings. A comparison with ten existing methods shows that our method outperforms the others in most cases.

Neural networks for NLP are becoming increasingly complex and widespread, and there is a growing concern if these models are responsible to use. Explaining models helps to address the safety and ethical concerns and is essential for accountability. Interpretability serves to provide these explanations in terms that are understandable to humans. Additionally, post-hoc methods provide explanations after a model is learned and are generally model-agnostic. This survey provides a categorization of how recent post-hoc interpretability methods communicate explanations to humans, it discusses each method in-depth, and how they are validated, as the latter is often a common concern.

Calabi-Yau four-folds may be constructed as hypersurfaces in weighted projective spaces of complex dimension 5 defined via weight systems of 6 weights. In this work, neural networks were implemented to learn the Calabi-Yau Hodge numbers from the weight systems, where gradient saliency and symbolic regression then inspired a truncation of the Landau-Ginzburg model formula for the Hodge numbers of any dimensional Calabi-Yau constructed in this way. The approximation always provides a tight lower bound, is shown to be dramatically quicker to compute (with compute times reduced by up to four orders of magnitude), and gives remarkably accurate results for systems with large weights. Additionally, complementary datasets of weight systems satisfying the necessary but insufficient conditions for transversality were constructed, including considerations of the IP, reflexivity, and intradivisibility properties. Overall producing a classification of this weight system landscape, further confirmed with machine learning methods. Using the knowledge of this classification, and the properties of the presented approximation, a novel dataset of transverse weight systems consisting of 7 weights was generated for a sum of weights $\leq 200$; producing a new database of Calabi-Yau five-folds, with their respective topological properties computed. Further to this an equivalent database of candidate Calabi-Yau six-folds was generated with approximated Hodge numbers.