Neural sequence generation models are known to "hallucinate", by producing outputs that are unrelated to the source text. These hallucinations are potentially harmful, yet it remains unclear in what conditions they arise and how to mitigate their impact. In this work, we first identify internal model symptoms of hallucinations by analyzing the relative token contributions to the generation in contrastive hallucinated vs. non-hallucinated outputs generated via source perturbations. We then show that these symptoms are reliable indicators of natural hallucinations, by using them to design a lightweight hallucination detector which outperforms both model-free baselines and strong classifiers based on quality estimation or large pre-trained models on manually annotated English-Chinese and German-English translation test beds.

We propose a way to improve the methodologies by introducing a new Language models (LMs) are ubiquitous in current score definition. During experiments we introduce NLP and have brought undeniable performance improvements several new variants of the existing datasets and for many tasks. Concerns have been a completely new dataset in Slovak. These new raised about the fairness of these models (Blodgett datasets are used to compare the expected behavior et al., 2020; Shah et al., 2020; Dev et al., 2021b). of the LMs with their actual behavior. Since LMs are usually trained with web-based text Our results challenge the validity of previous corpora generated by a general population, there studies.

The issue of fairness in AI has received an increasing amount of attention in recent years. The problem can be approached by looking at different protected attributes (e.g., ethnicity, gender, etc) independently, but fairness for individual protected attributes does not imply intersectional fairness. In this work, we frame the problem of intersectional fairness within a geometrical setting. We project our data onto a hypercube, and split the analysis of fairness by levels, where each level encodes the number of protected attributes we are intersecting over. We prove mathematically that, while fairness does not propagate "down" the levels, it does propagate "up" the levels. This means that ensuring fairness for all subgroups at the lowest intersectional level (e.g., black women, white women, black men and white men), will necessarily result in fairness for all the above levels, including each of the protected attributes (e.g., ethnicity and gender) taken independently. We also derive a formula describing the variance of the set of estimated success rates on each level, under the assumption of perfect fairness. Using this theoretical finding as a benchmark, we define a family of metrics which capture overall intersectional bias. Finally, we propose that fairness can be metaphorically thought of as a "fractal" problem. In fractals, patterns at the smallest scale repeat at a larger scale. We see from this example that tackling the problem at the lowest possible level, in a bottom-up manner, leads to the natural emergence of fair AI. We suggest that trustworthiness is necessarily an emergent, fractal and relational property of the AI system.

Variational Autoencoder (VAE)-based generative models offer flexible representation learning by incorporating meta-priors, general premises considered beneficial for downstream tasks. However, the incorporated meta-priors often involve ad-hoc model deviations from the original likelihood architecture, causing undesirable changes in their training. In this paper, we propose a novel representation learning method, Gromov-Wasserstein Autoencoders (GWAE), which directly matches the latent and data distributions using the variational autoencoding scheme. Instead of likelihood-based objectives, GWAE models minimize the Gromov-Wasserstein (GW) metric between the trainable prior and given data distributions. The GW metric measures the distance structure-oriented discrepancy between distributions even with different dimensionalities, which provides a direct measure between the latent and data spaces. By restricting the prior family, we can introduce meta-priors into the latent space without changing their objective. The empirical comparisons with VAE-based models show that GWAE models work in two prominent meta-priors, disentanglement and clustering, with their GW objective unchanged.

Training deep neural networks (DNNs) usually requires massive training data and computational resources. Users who cannot afford this may prefer to outsource training to a third party or resort to publicly available pre-trained models. Unfortunately, doing so facilitates a new training-time attack (i.e., backdoor attack) against DNNs. This attack aims to induce misclassification of input samples containing adversary-specified trigger patterns. In this paper, we first conduct a layer-wise feature analysis of poisoned and benign samples from the target class. We find out that the feature difference between benign and poisoned samples tends to be maximum at a critical layer, which is not always the one typically used in existing defenses, namely the layer before fully-connected layers. We also demonstrate how to locate this critical layer based on the behaviors of benign samples. We then propose a simple yet effective method to filter poisoned samples by analyzing the feature differences between suspicious and benign samples at the critical layer. We conduct extensive experiments on two benchmark datasets, which confirm the effectiveness of our defense.

Internet censorship is a phenomenon of societal importance and attracts investigation from multiple disciplines. Several research groups, such as Censored Planet, have deployed large scale Internet measurement platforms to collect network reachability data. However, existing studies generally rely on manually designed rules (i.e., using censorship fingerprints) to detect network-based Internet censorship from the data. While this rule-based approach yields a high true positive detection rate, it suffers from several challenges: it requires human expertise, is laborious, and cannot detect any censorship not captured by the rules. Seeking to overcome these challenges, we design and evaluate a classification model based on latent feature representation learning and an image-based classification model to detect network-based Internet censorship. To infer latent feature representations fromnetwork reachability data, we propose a sequence-to-sequence autoencoder to capture the structure and the order of data elements in the data. To estimate the probability of censorship events from the inferred latent features, we rely on a densely connected multi-layer neural network model. Our image-based classification model encodes a network reachability data record as a gray-scale image and classifies the image as censored or not using a dense convolutional neural network. We compare and evaluate both approaches using data sets from Censored Planet via a hold-out evaluation. Both classification models are capable of detecting network-based Internet censorship as we were able to identify instances of censorship not detected by the known fingerprints. Latent feature representations likely encode more nuances in the data since the latent feature learning approach discovers a greater quantity, and a more diverse set, of new censorship instances.

Since their proposal in the 2014 paper by Ian Goodfellow, there has been an explosion of research into the area of Generative Adversarial Networks. While they have been utilised in many fields, the realm of malware research is a problem space in which GANs have taken root. From balancing datasets to creating unseen examples in rare classes, GAN models offer extensive opportunities for application. This paper surveys the current research and literature for the use of Generative Adversarial Networks in the malware problem space. This is done with the hope that the reader may be able to gain an overall understanding as to what the Generative Adversarial model provides for this field, and for what areas within malware research it is best utilised. It covers the current related surveys, the different categories of GAN, and gives the outcomes of recent research into optimising GANs for different topics, as well as future directions for exploration.

We study the statistical properties of learning to defer (L2D) to multiple experts. In particular, we address the open problems of deriving a consistent surrogate loss, confidence calibration, and principled ensembling of experts. Firstly, we derive two consistent surrogates -- one based on a softmax parameterization, the other on a one-vs-all (OvA) parameterization -- that are analogous to the single expert losses proposed by Mozannar and Sontag (2020) and Verma and Nalisnick (2022), respectively. We then study the frameworks' ability to estimate P( m_j = y | x ), the probability that the jth expert will correctly predict the label for x. Theory shows the softmax-based loss causes mis-calibration to propagate between the estimates while the OvA-based loss does not (though in practice, we find there are trade offs). Lastly, we propose a conformal inference technique that chooses a subset of experts to query when the system defers. We perform empirical validation on tasks for galaxy, skin lesion, and hate speech classification.

This paper studies algorithmic fairness when the protected attribute is location. To handle protected attributes that are continuous, such as age or income, the standard approach is to discretize the domain into predefined groups, and compare algorithmic outcomes across groups. However, applying this idea to location raises concerns of gerrymandering and may introduce statistical bias. Prior work addresses these concerns but only for regularly spaced locations, while raising other issues, most notably its inability to discern regions that are likely to exhibit spatial unfairness. Similar to established notions of algorithmic fairness, we define spatial fairness as the statistical independence of outcomes from location. This translates into requiring that for each region of space, the distribution of outcomes is identical inside and outside the region. To allow for localized discrepancies in the distribution of outcomes, we compare how well two competing hypotheses explain the observed outcomes. The null hypothesis assumes spatial fairness, while the alternate allows different distributions inside and outside regions. Their goodness of fit is then assessed by a likelihood ratio test. If there is no significant difference in how well the two hypotheses explain the observed outcomes, we conclude that the algorithm is spatially fair.

Accurately estimating the shape of objects in dense clutters makes important contribution to robotic packing, because the optimal object arrangement requires the robot planner to acquire shape information of all existed objects. However, the objects for packing are usually piled in dense clutters with severe occlusion, and the object shape varies significantly across different instances for the same category. They respectively cause large object segmentation errors and inaccurate shape recovery on unseen instances, which both degrade the performance of shape estimation during deployment. In this paper, we propose a category-level shape estimation method for densely cluttered objects. Our framework partitions each object in the clutter via the multi-view visual information fusion to achieve high segmentation accuracy, and the instance shape is recovered by deforming the category templates with diverse geometric transformations to obtain strengthened generalization ability. Specifically, we first collect the multi-view RGB-D images of the object clutters for point cloud reconstruction. Then we fuse the feature maps representing the visual information of multi-view RGB images and the pixel affinity learned from the clutter point cloud, where the acquired instance segmentation masks of multi-view RGB images are projected to partition the clutter point cloud. Finally, the instance geometry information is obtained from the partially observed instance point cloud and the corresponding category template, and the deformation parameters regarding the template are predicted for shape estimation. Experiments in the simulated environment and real world show that our method achieves high shape estimation accuracy for densely cluttered everyday objects with various shapes.