Search

There is a dire need for medical imaging datasets with accompanying annotations to perform downstream patient analysis. However, it is difficult to manually generate these annotations, due to the time-consuming nature, and the variability in clinical conventions. Artificial intelligence has been adopted in the field as a potential method to annotate these large datasets, however, a lack of expert annotations or ground truth can inhibit the adoption of these annotations. We recently made a dataset publicly available including annotations and extracted features of up to 104 organs for the National Lung Screening Trial using the TotalSegmentator method. However, the released dataset does not include expert-derived annotations or an assessment of the accuracy of the segmentations, limiting its usefulness. We propose the development of heuristics to assess the quality of the segmentations, providing methods to measure the consistency of the annotations and a comparison of results to the literature. We make our code and related materials publicly available at https://github.com/ImagingDataCommons/CloudSegmentatorResults and interactive tools at https://huggingface.co/spaces/ImagingDataCommons/CloudSegmentatorResults.

This paper explores the improvement of post-training quantization (PTQ) after knowledge distillation in the Whisper speech foundation model family. We address the challenge of outliers in weights and activation tensors, known to impede quantization quality in transformer-based language and vision models. Extending this observation to Whisper, we demonstrate that these outliers are also present when transformer-based models are trained to perform automatic speech recognition, necessitating mitigation strategies for PTQ. We show that outliers can be reduced by a recently proposed gating mechanism in the attention blocks of the student model, enabling effective 8-bit quantization, and lower word error rates compared to student models without the gating mechanism in place., Comment: Accepted at Interspeech 2024

A methodology based on deep recurrent models for maritime surveillance, over publicly available Automatic Identification System (AIS) data, is presented in this paper. The setup employs a deep Recurrent Neural Network (RNN)-based model, for encoding and reconstructing the observed ships' motion patterns. Our approach is based on a thresholding mechanism, over the calculated errors between observed and reconstructed motion patterns of maritime vessels. Specifically, a deep-learning framework, i.e. an encoder-decoder architecture, is trained using the observed motion patterns, enabling the models to learn and predict the expected trajectory, which will be compared to the effective ones. Our models, particularly the bidirectional GRU with recurrent dropouts, showcased superior performance in capturing the temporal dynamics of maritime data, illustrating the potential of deep learning to enhance maritime surveillance capabilities. Our work lays a solid foundation for future research in this domain, highlighting a path toward improved maritime safety through the innovative application of technology., Comment: Presented in PETRA '24 The PErvasive Technologies Related to Assistive Environments Conference June 26--28, 2024 Crete, Greece

Random effects meta-analysis model is an important tool for integrating results from multiple independent studies. However, the standard model is based on the assumption of normal distributions for both random effects and within-study errors, making it susceptible to outlying studies. Although robust modeling using the $t$ distribution is an appealing idea, the existing work, that explores the use of the $t$ distribution only for random effects, involves complicated numerical integration and numerical optimization. In this paper, a novel robust meta-analysis model using the $t$ distribution is proposed ($t$Meta). The novelty is that the marginal distribution of the effect size in $t$Meta follows the $t$ distribution, enabling that $t$Meta can simultaneously accommodate and detect outlying studies in a simple and adaptive manner. A simple and fast EM-type algorithm is developed for maximum likelihood estimation. Due to the mathematical tractability of the $t$ distribution, $t$Meta frees from numerical integration and allows for efficient optimization. Experiments on real data demonstrate that $t$Meta is compared favorably with related competitors in situations involving mild outliers. Moreover, in the presence of gross outliers, while related competitors may fail, $t$Meta continues to perform consistently and robustly., Comment: 15 pages, 7 figures

In this paper, we propose a novel conceptual framework to detect outliers using optimal transport with a concave cost function. Conventional outlier detection approaches typically use a two-stage procedure: first, outliers are detected and removed, and then estimation is performed on the cleaned data. However, this approach does not inform outlier removal with the estimation task, leaving room for improvement. To address this limitation, we propose an automatic outlier rectification mechanism that integrates rectification and estimation within a joint optimization framework. We take the first step to utilize the optimal transport distance with a concave cost function to construct a rectification set in the space of probability distributions. Then, we select the best distribution within the rectification set to perform the estimation task. Notably, the concave cost function we introduced in this paper is the key to making our estimator effectively identify the outlier during the optimization process. We demonstrate the effectiveness of our approach over conventional approaches in simulations and empirical analyses for mean estimation, least absolute regression, and the fitting of option implied volatility surfaces.

Explaining outliers occurrence and mechanism of their occurrence can be extremely important in a variety of domains. Malfunctions, frauds, threats, in addition to being correctly identified, oftentimes need a valid explanation in order to effectively perform actionable counteracts. The ever more widespread use of sophisticated Machine Learning approach to identify anomalies make such explanations more challenging. We present the Decision Tree Outlier Regressor (DTOR), a technique for producing rule-based explanations for individual data points by estimating anomaly scores generated by an anomaly detection model. This is accomplished by first applying a Decision Tree Regressor, which computes the estimation score, and then extracting the relative path associated with the data point score. Our results demonstrate the robustness of DTOR even in datasets with a large number of features. Additionally, in contrast to other rule-based approaches, the generated rules are consistently satisfied by the points to be explained. Furthermore, our evaluation metrics indicate comparable performance to Anchors in outlier explanation tasks, with reduced execution time.

This work introduces the Matrix Minimum Covariance Determinant (MMCD) method, a novel robust location and covariance estimation procedure designed for data that are naturally represented in the form of a matrix. Unlike standard robust multivariate estimators, which would only be applicable after a vectorization of the matrix-variate samples leading to high-dimensional datasets, the MMCD estimators account for the matrix-variate data structure and consistently estimate the mean matrix, as well as the rowwise and columnwise covariance matrices in the class of matrix-variate elliptical distributions. Additionally, we show that the MMCD estimators are matrix affine equivariant and achieve a higher breakdown point than the maximal achievable one by any multivariate, affine equivariant location/covariance estimator when applied to the vectorized data. An efficient algorithm with convergence guarantees is proposed and implemented. As a result, robust Mahalanobis distances based on MMCD estimators offer a reliable tool for outlier detection. Additionally, we extend the concept of Shapley values for outlier explanation to the matrix-variate setting, enabling the decomposition of the squared Mahalanobis distances into contributions of the rows, columns, or individual cells of matrix-valued observations. Notably, both the theoretical guarantees and simulations show that the MMCD estimators outperform robust estimators based on vectorized observations, offering better computational efficiency and improved robustness. Moreover, real-world data examples demonstrate the practical relevance of the MMCD estimators and the resulting robust Shapley values.

Outlier detection is an essential capability in safety-critical applications of supervised visual recognition. Most of the existing methods deliver best results by encouraging standard closed-set models to produce low-confidence predictions in negative training data. However, that approach conflates prediction uncertainty with recognition of the negative class. We therefore reconsider direct prediction of K+1 logits that correspond to K groundtruth classes and one outlier class. This setup allows us to formulate a novel anomaly score as an ensemble of in-distribution uncertainty and the posterior of the outlier class which we term negative objectness. Now outliers can be independently detected due to i) high prediction uncertainty or ii) similarity with negative data. We embed our method into a dense prediction architecture with mask-level recognition over K+2 classes. The training procedure encourages the novel K+2-th class to learn negative objectness at pasted negative instances. Our models outperform the current state-of-the art on standard benchmarks for image-wide and pixel-level outlier detection with and without training on real negative data., Comment: Accepted to BMVC 2024

Mass lumping techniques are commonly employed in explicit time integration schemes for problems in structural dynamics and both avoid solving costly linear systems with the consistent mass matrix and increase the critical time step. In isogeometric analysis, the critical time step is constrained by so-called "outlier" frequencies, representing the inaccurate high frequency part of the spectrum. Removing or dampening these high frequencies is paramount for fast explicit solution techniques. In this work, we propose robust mass lumping and outlier removal techniques for nontrivial geometries, including multipatch and trimmed geometries. Our lumping strategies provably do not deteriorate (and often improve) the CFL condition of the original problem and are combined with deflation techniques to remove persistent outlier frequencies. Numerical experiments reveal the advantages of the method, especially for simulations covering large time spans where they may halve the number of iterations with little or no effect on the numerical solution., Comment: 31 pages, 16 figures. Submitted manuscript

Outlier generation is a popular technique used for solving important outlier detection tasks. Generating outliers with realistic behavior is challenging. Popular existing methods tend to disregard the 'multiple views' property of outliers in high-dimensional spaces. The only existing method accounting for this property falls short in efficiency and effectiveness. We propose BISECT, a new outlier generation method that creates realistic outliers mimicking said property. To do so, BISECT employs a novel proposition introduced in this article stating how to efficiently generate said realistic outliers. Our method has better guarantees and complexity than the current methodology for recreating 'multiple views'. We use the synthetic outliers generated by BISECT to effectively enhance outlier detection in diverse datasets, for multiple use cases. For instance, oversampling with BISECT reduced the error by up to 3 times when compared with the baselines., Comment: Preprint. Full paper is scheduled to appear in TKDD; Updated results in table 4

The Minimum Covariance Determinant (MCD) method is a widely adopted tool for robust estimation and outlier detection. In this paper, we introduce a new framework for model selection in MCD with spectral embedding based on the notion of stability. Our best subset algorithm leverages principal component analysis for dimension reduction, statistical depths for effective initialization, and concentration steps for subset refinement. Subsequently, we construct a bootstrap procedure to estimate the instability of the best subset algorithm. The parameter combination exhibiting minimal instability proves ideal for the purposes of high-dimensional outlier detection, while the instability path offers insights into the inlier/outlier structure. We rigorously benchmark the proposed framework against existing MCD variants and illustrate its practical utility on two spectra data sets and a cancer genomics data set.

We present a nonparametric method for outlier detection that takes full account of local variations in intrinsic dimensionality within the dataset. Using the theory of Local Intrinsic Dimensionality (LID), our 'dimensionality-aware' outlier detection method, DAO, is derived as an estimator of an asymptotic local expected density ratio involving the query point and a close neighbor drawn at random. The dimensionality-aware behavior of DAO is due to its use of local estimation of LID values in a theoretically-justified way. Through comprehensive experimentation on more than 800 synthetic and real datasets, we show that DAO significantly outperforms three popular and important benchmark outlier detection methods: Local Outlier Factor (LOF), Simplified LOF, and kNN., Comment: 13 pages, 3 figures. Extended version of a paper accepted for publication at the SIAM International Conference on Data Mining (SDM24)

Defining a successful notion of a multivariate quantile has been an open problem for more than half a century, motivating a plethora of possible solutions. Of these, the approach of [8] and [25] leading to M-quantiles, is very appealing for its mathematical elegance combining elements of convex analysis and probability theory. The key idea is the description of a convex function (the K-function) whose gradient (the K-transform) is in one-to-one correspondence between all of R^d and the unit ball in R^d. By analogy with the d=1 case where the K-transform is a cumulative distribution function-like object (an M-distribution), the fact that its inverse is guaranteed to exist lends itself naturally to providing the basis for the definition of a quantile function for all d>=1. Over the past twenty years the resulting M-quantiles have seen applications in a variety of fields, primarily for the purpose of detecting outliers in multidimensional spaces. In this article we prove that for odd d>=3, it is not the gradient but a poly-Laplacian of the K-function that is (almost everywhere) proportional to the density function. For d even one cannot establish a differential equation connecting the K-function with the density. These results show that usage of the K-transform for outlier detection in higher odd-dimensions is in principle flawed, as the K-transform does not originate from inversion of a true M-distribution. We demonstrate these conclusions in two dimensions through examples from non-standard asymmetric distributions. Our examples illustrate a feature of the K-transform whereby regions in the domain with higher density map to larger volumes in the co-domain, thereby producing a magnification effect that moves inliers closer to the boundary of the co-domain than outliers. This feature obviously disrupts any outlier detection mechanism that relies on the inverse K-transform.

Disease control experts inspect public health data streams daily for outliers worth investigating, like those corresponding to data quality issues or disease outbreaks. However, they can only examine a few of the thousands of maximally-tied outliers returned by univariate outlier detection methods applied to large-scale public health data streams. To help experts distinguish the most important outliers from these thousands of tied outliers, we propose a new task for algorithms to rank the outputs of any univariate method applied to each of many streams. Our novel algorithm for this task, which leverages hierarchical networks and extreme value analysis, performed the best across traditional outlier detection metrics in a human-expert evaluation using public health data streams. Most importantly, experts have used our open-source Python implementation since April 2023 and report identifying outliers worth investigating 9.1x faster than their prior baseline. Other organizations can readily adapt this implementation to create rankings from the outputs of their tailored univariate methods across large-scale streams., Comment: 6 figures, 8 pages

Probabilistic mixture models are recognized as effective tools for unsupervised outlier detection owing to their interpretability and global characteristics. Among these, Dirichlet process mixture models stand out as a strong alternative to conventional finite mixture models for both clustering and outlier detection tasks. Unlike finite mixture models, Dirichlet process mixtures are infinite mixture models that automatically determine the number of mixture components based on the data. Despite their advantages, the adoption of Dirichlet process mixture models for unsupervised outlier detection has been limited by challenges related to computational inefficiency and sensitivity to outliers in the construction of outlier detectors. Additionally, Dirichlet process Gaussian mixtures struggle to effectively model non-Gaussian data with discrete or binary features. To address these challenges, we propose a novel outlier detection method that utilizes ensembles of Dirichlet process Gaussian mixtures. This unsupervised algorithm employs random subspace and subsampling ensembles to ensure efficient computation and improve the robustness of the outlier detector. The ensemble approach further improves the suitability of the proposed method for detecting outliers in non-Gaussian data. Furthermore, our method uses variational inference for Dirichlet process mixtures, which ensures both efficient and rapid computation. Empirical analyses using benchmark datasets demonstrate that our method outperforms existing approaches in unsupervised outlier detection.

Race conditions are a class of bugs in software where concurrent accesses to shared resources are not protected from each other. Consequences of race conditions include privilege escalation, denial of service, and memory corruption which can potentially lead to arbitrary code execution. However, in large code bases the exact rules as to which fields should be accessed under which locks are not always clear. We propose a novel static technique that infers rules for how field accesses should be locked, and then checks the code against these rules. Traditional static analysers for detecting race conditions are based on lockset analysis. Instead, we propose an outlier-based technique enhanced with a context-sensitive mechanism that scales well. We have implemented this analysis in LLIF, and evaluated it to find incorrectly protected field accesses in Linux v5.14.11. We thoroughly evaluate its ability to find race conditions, and study the causes for false positive reports. In addition, we reported a subset of the issues and submitted patches. The maintainers confirmed 24 bugs.

We study the problem of robust multivariate polynomial regression: let $p\colon\mathbb{R}^n\to\mathbb{R}$ be an unknown $n$-variate polynomial of degree at most $d$ in each variable. We are given as input a set of random samples $(\mathbf{x}_i,y_i) \in [-1,1]^n \times \mathbb{R}$ that are noisy versions of $(\mathbf{x}_i,p(\mathbf{x}_i))$. More precisely, each $\mathbf{x}_i$ is sampled independently from some distribution $\chi$ on $[-1,1]^n$, and for each $i$ independently, $y_i$ is arbitrary (i.e., an outlier) with probability at most $\rho < 1/2$, and otherwise satisfies $|y_i-p(\mathbf{x}_i)|\leq\sigma$. The goal is to output a polynomial $\hat{p}$, of degree at most $d$ in each variable, within an $\ell_\infty$-distance of at most $O(\sigma)$ from $p$. Kane, Karmalkar, and Price [FOCS'17] solved this problem for $n=1$. We generalize their results to the $n$-variate setting, showing an algorithm that achieves a sample complexity of $O_n(d^n\log d)$, where the hidden constant depends on $n$, if $\chi$ is the $n$-dimensional Chebyshev distribution. The sample complexity is $O_n(d^{2n}\log d)$, if the samples are drawn from the uniform distribution instead. The approximation error is guaranteed to be at most $O(\sigma)$, and the run-time depends on $\log(1/\sigma)$. In the setting where each $\mathbf{x}_i$ and $y_i$ are known up to $N$ bits of precision, the run-time's dependence on $N$ is linear. We also show that our sample complexities are optimal in terms of $d^n$. Furthermore, we show that it is possible to have the run-time be independent of $1/\sigma$, at the cost of a higher sample complexity.

Transformer-based models have gained widespread popularity in both the computer vision (CV) and natural language processing (NLP) fields. However, significant challenges arise during post-training linear quantization, leading to noticeable reductions in inference accuracy. Our study focuses on uncovering the underlying causes of these accuracy drops and proposing a quantization-friendly fine-tuning method, \textbf{QuantTune}. Firstly, our analysis revealed that, on average, 65\% of quantization errors result from the precision loss incurred by the dynamic range amplification effect of outliers across the target Transformer-based models. Secondly, \textbf{QuantTune} adjusts weights based on the deviation of outlier activations and effectively constrains the dynamic ranges of the problematic activations. As a result, it successfully mitigates the negative impact of outliers on the inference accuracy of quantized models. Lastly, \textbf{QuantTune} can be seamlessly integrated into the back-propagation pass in the fine-tuning process without requiring extra complexity in inference software and hardware design. Our approach showcases significant improvements in post-training quantization across a range of Transformer-based models, including ViT, Bert-base, and OPT. QuantTune reduces accuracy drops by 12.09\% at 8-bit quantization and 33.8\% at 7-bit compared to top calibration methods, outperforming state-of-the-art solutions by over 18.84\% across ViT models.

Atom probe tomography is commonly used to study solute clustering and precipitation in materials. However, standard techniques, such as the density based spatial clustering applications with noise (DBSCAN) perform poorly with respect to small clusters of less than 25 atoms. This is a fundamental limitation of density-based clustering techniques due to the usage of Nmin, an arbitrary lower limit placed on cluster sizes. Therefore, this paper attempts to consider atom probe clustering as an outlier detection problem of which KNN, LOF, LUNAR algorithms were tested against a simulated dataset and compared to the standard method, for a range of cluster sizes. The decision score output of the algorithms was then auto thresholded by the Karcher mean to remove human bias. Each of the major models tested outperforms DBSCAN for cluster sizes of less than 25 atoms but underperforms for sizes greater than 30 atoms. However, the new combined kNN and DBSCAN method presented was able to perform well at all cluster sizes. The combined kNN and DBSCAN method is presented as a possible new standard approach to identifying solute clusters in atom probe tomography.

Lightweight design of Convolutional Neural Networks (CNNs) requires co-design efforts in the model architectures and compression techniques. As a novel design paradigm that separates training and inference, a structural re-parameterized (SR) network such as the representative RepVGG revitalizes the simple VGG-like network with a high accuracy comparable to advanced and often more complicated networks. However, the merging process in SR networks introduces outliers into weights, making their distribution distinct from conventional networks and thus heightening difficulties in quantization. To address this, we propose an operator-level improvement for training called Outlier Aware Batch Normalization (OABN). Additionally, to meet the demands of limited bitwidths while upkeeping the inference accuracy, we develop a clustering-based non-uniform quantization framework for Quantization-Aware Training (QAT) named ClusterQAT. Integrating OABN with ClusterQAT, the quantized performance of RepVGG is largely enhanced, particularly when the bitwidth falls below 8., Comment: 8 pages, 8 figures

Reliable and precise absolute positioning is necessary in the realm of Connected Automated Vehicles (CAV). Global Navigation Satellite Systems (GNSS) provides the foundation for absolute positioning. Recently enhanced Precise Point Positioning (PPP) technology now offers corrections for GNSS on a global scale, with the potential to achieve accuracy suitable for real-time CAV applications. However, in obstructed sky conditions, GNSS signals are often affected by outliers; therefore, addressing outliers is crucial. In GNSS applications, there are many more measurements available than are required to meet the specification. Therefore, selecting measurements to avoid outliers is of interest. The recently developed Risk-Averse Performance-Specified (RAPS) state estimation optimally selects measurements to minimize outlier risk while meeting a positive semi-definite constraint on performance; at present, the existing solution methods are not suitable for real-time computation and have not been demonstrated using challenging real-world data or in Real-time PPP (RT-PPP) applications. This article makes contributions in a few directions. First, it uses a diagonal performance specification, which reduces computational costs relative to the positive semi-definite constraint. Second, this article considers GNSS RT-PPP applications. Third, the experiments use real-world GNSS data collected in challenging environments. The RT-PPP experimental results show that among the compared methods: all achieve comparable performance in open-sky conditions, and all exceed the Society of Automotive Engineers (SAE) specification; however, in challenging environments, the diagonal RAPS approach shows improvement of 6-19% over traditional methods. Throughout, RAPS achieves the lowest estimation risk., Comment: 7 pages,2 figures, Accepted by 2024 American Control Conference

We introduce QuaRot, a new Quantization scheme based on Rotations, which is able to quantize LLMs end-to-end, including all weights, activations, and KV cache in 4 bits. QuaRot rotates LLMs in a way that removes outliers from the hidden state without changing the output, making quantization easier. This computational invariance is applied to the hidden state (residual) of the LLM, as well as to the activations of the feed-forward components, aspects of the attention mechanism and to the KV cache. The result is a quantized model where all matrix multiplications are performed in 4-bits, without any channels identified for retention in higher precision. Our quantized LLaMa2-70B model has losses of at most 0.29 WikiText-2 perplexity and retains 99% of the zero-shot performance. Code is available at: https://github.com/spcl/QuaRot., Comment: 19 pages, 6 figures

Out-Of-Distribution (OOD) detection is critical to deploy deep learning models in safety-critical applications. However, the inherent hierarchical concept structure of visual data, which is instrumental to OOD detection, is often poorly captured by conventional methods based on Euclidean geometry. This work proposes a metric framework that leverages the strengths of Hyperbolic geometry for OOD detection. Inspired by previous works that refine the decision boundary for OOD data with synthetic outliers, we extend this method to Hyperbolic space. Interestingly, we find that synthetic outliers do not benefit OOD detection in Hyperbolic space as they do in Euclidean space. Furthermore we explore the relationship between OOD detection performance and Hyperbolic embedding dimension, addressing practical concerns in resource-constrained environments. Extensive experiments show that our framework improves the FPR95 for OOD detection from 22\% to 15\% and from 49% to 28% on CIFAR-10 and CIFAR-100 respectively compared to Euclidean methods.

Uncertainty quantification (UQ) to detect samples with large expected errors (outliers) is applied to reactive molecular potential energy surfaces (PESs). Three methods - Ensembles, Deep Evidential Regression (DER), and Gaussian Mixture Models (GMM) - were applied to the H-transfer reaction between ${\it syn-}$Criegee and vinyl hydroxyperoxide. The results indicate that ensemble models provide the best results for detecting outliers, followed by GMM. For example, from a pool of 1000 structures with the largest uncertainty, the detection quality for outliers is $\sim 90$ \% and $\sim 50$ \%, respectively, if 25 or 1000 structures with large errors are sought. On the contrary, the limitations of the statistical assumptions of DER greatly impacted its prediction capabilities. Finally, a structure-based indicator was found to be correlated with large average error, which may help to rapidly classify new structures into those that provide an advantage for refining the neural network.

Recent works have shown that by using large pre-trained models along with learnable prompts, rehearsal-free methods for class-incremental learning (CIL) settings can achieve superior performance to prominent rehearsal-based ones. Rehearsal-free CIL methods struggle with distinguishing classes from different tasks, as those are not trained together. In this work we propose a regularization method based on virtual outliers to tighten decision boundaries of the classifier, such that confusion of classes among different tasks is mitigated. Recent prompt-based methods often require a pool of task-specific prompts, in order to prevent overwriting knowledge of previous tasks with that of the new task, leading to extra computation in querying and composing an appropriate prompt from the pool. This additional cost can be eliminated, without sacrificing accuracy, as we reveal in the paper. We illustrate that a simplified prompt-based method can achieve results comparable to previous state-of-the-art (SOTA) methods equipped with a prompt pool, using much less learnable parameters and lower inference cost. Our regularization method has demonstrated its compatibility with different prompt-based methods, boosting those previous SOTA rehearsal-free CIL methods' accuracy on the ImageNet-R and CIFAR-100 benchmarks. Our source code is available at https://github.com/jpmorganchase/ovor., Comment: Accepted by ICLR 2024