The paper corrects biases in estimating intrinsic dimension and differential entropy.
problem Systematic bias in estimating intrinsic dimension and differential entropy.
method A bias-corrected estimator for both measures is proposed, highlighting shared steps and useful consequences.
result Simultaneous estimation of differential entropy and intrinsic dimension provides complementary perspectives on underlying manifolds.
It is commonly believed that increasing the interpretability of a machine learning model may decrease its predictive power. However, inspecting input-output relationships of those models using visual analytics, while treating them as black-box, can help to understand the reasoning behind outcomes without sacrificing pr…
Machine learning analysis of galaxy catalogues finds only one class separable.
problem Difficulty in classifying galaxies using visual inspection schemes.
method Generalized Relevance Matrix Learning Vector Quantization and Random Forests.
result Only one class, Little Blue Spheroids, is consistently separable.
A new system detects and classifies defects in semiconductor manufacturing.
problem Detecting and classifying novel defect patterns in high-resolution imagery.
method Stacked hybrid convolutional neural networks (SH-CNN) with visual attention.
result SH-CNN outperforms current approaches in automated visual inspection.
Study compares deep feature methods for anomaly detection in limited data scenarios.
problem Handling limited data in industrial inspection applications.
method Three approaches (KNN, Mahalanobis, PaDiM) using pre-trained deep features with data augmentation.
result Data augmentation significantly improves performance in small data regimes.
Bayes-TrEx finds in-distribution examples for model inspection.
problem Challenges in interpreting neural networks, especially high-confidence failures and ambiguous classifications.
method Bayesian sampling approach to find in-distribution examples with specified prediction confidence.
result Bayes-TrEx enables more flexible holistic model analysis than just inspecting the test set.
Active learning improves inspection systems by using weakly labeled data.
problem Rapidly updating machine vision inspection systems in evolving manufacturing processes.
method Developed a methodology for active learning from weakly labeled data, addressing covariate shift with domain-adversarial training.
result Demonstrated that active learning can accelerate the annotation process and reduce false positives.
DecompKAN improves time series forecasting accuracy and transparency.
problem Accurate and transparent time series forecasting in scientific domains.
method Combines decomposition, patching, normalization, and B-spline KAN edge functions.
result Achieves best or tied-best MSE on 20 of 36 comparisons across 9 datasets.
A network removes irrelevant structures from chest radiographs for better analysis.
problem Clutter in chest radiographs hinders visual inspection and analysis.
method Fully Convolutional Network to suppress undesired visual structure.
result Improved classifier performance with limited training data.
Paper proposes using NLPD for cGANs to improve image realism and segmentation.
problem Lack of perceptual evaluation in cGANs.
method Used NLPD to measure perceptual similarity, comparing with L1 distance.
result NLPD improves realism and segmentation accuracy in generated images.
New method evaluates visual explanations of deep models using adversarial perturbations.
problem Lack of objective evaluation of visual explanations of deep models.
method Proposes an adversarial perturbation approach to evaluate visual explanations of deep models.
result Demonstrates the effectiveness of the proposed approach through comparisons with existing methods.
Although deep learning techniques have been successfully applied to many tasks, interpreting deep neural network models is still a big challenge to us. Recently, many works have been done on visualizing and analyzing the mechanism of deep neural networks in the areas of image processing and natural language processing.…
ConfusionFlow visualizes classifier confusion over time for model comparison.
problem Insufficient performance analysis of classifiers.
method Interactive, model-agnostic visualization tool combining confusion matrices and temporal analysis.
result ConfusionFlow facilitates detailed, comparative analysis of classifier performance over time.
Using different methods for laying out a graph can lead to very different visual appearances, with which the viewer perceives different information. Selecting a "good" layout method is thus important for visualizing a graph. The selection can be highly subjective and dependent on the given task. A common approach to se…
Visualizes deep neural network decisions in 2D for better understanding.
problem Difficulty in comprehending complex deep learning models.
method Discriminative dimensionality reduction for 2D visualization of model decisions.
result Insight into how different data properties are treated by the model.
AttViz offers online visualizations of neural language model attention mechanisms.
problem Limited interpretability of neural language models.
method Online toolkit for exploring self-attention mechanisms.
result Visualizations help understand model decision-making.
Proposes manifold-based unsupervised anomaly detection for visual data.
problem Rare anomalies in unlabeled data.
method Constant curvature manifolds, hyperspherical Variational Auto-Encoders (VAE) with gyroplane layer.
result State-of-the-art results on visual anomaly benchmarks and histopathology.
Semi-automatic data annotation helps experts label unlabeled samples based on feature space projection.
problem Laborious manual data annotation for machine learning.
method Interactive semi-automatic approach using feature space projection and semi-supervised learning.
result Reduces user annotation effort and improves classification accuracy.
Generating molecules with desired chemical properties is important for drug discovery. The use of generative neural networks is promising for this task. However, from visual inspection, it often appears that generated samples lack diversity. In this paper, we quantify this internal chemical diversity, and we raise the …
Interactive tool for better understanding t-SNE projections.
problem Interpreting t-SNE projections can be challenging and misleading.
method Interactive visualization tool with different views.
result Improves understanding of t-SNE and its results.
New method uses DTW to evaluate neural network forecasts of geomagnetic indices.
problem Evaluation metrics fail to capture persistence behavior in neural network forecasts.
method Dynamic Time Warping (DTW) to measure time series similarity, training neural networks to remove persistence.
result DTW reveals persistence behavior in neural network forecasts, confirming visual inspection.
Interpretation and diagnosis of machine learning models have gained renewed interest in recent years with breakthroughs in new approaches. We present Manifold, a framework that utilizes visual analysis techniques to support interpretation, debugging, and comparison of machine learning models in a more transparent and i…
New tool for interpreting neural nets, explaining model behavior.
problem Interpreting neural network responses and explaining model behavior.
method Full-Gradient representation and FullGrad saliency map approximation.
result FullGrad method explains model behavior more comprehensively and accurately than other methods.
Availability of an explainable deep learning model that can be applied to practical real world scenarios and in turn, can consistently, rapidly and accurately identify specific and minute traits in applicable fields of biological sciences, is scarce. Here we consider one such real world example viz., accurate identific…
Framework for efficient defect classification and inspection.
problem Adaptive defect classification and inspection from high volume data.
method Continual learning framework for dynamic classifier updates.
result Efficient storage and computational needs reduction.
VisRuler simplifies decision extraction from bagged and boosted trees.
problem Complexity and lack of interpretability in ensemble models.
method Visual analytics tool for selecting robust models, important features, and essential decisions.
result Users successfully extracted and explained decisions from ensemble models.
ChainerRL is a deep reinforcement learning library for Python.
problem Training and reproducing deep reinforcement learning algorithms.
method Built on Chainer framework, implements various DRL algorithms.
result Fosters reproducible research and replicates benchmark results.
Study Fisher information for detecting neural network adversarial attacks.
problem Detecting and understanding adversarial examples in neural networks.
method Compute Fisher information quantities efficiently and analyze their behavior on adversarial examples.
result Fisher information can highlight important input neurons and reveal unreasonable network behavior.
The paper introduces closed-form expressions for interpreting Tsetlin Machines.
problem Interpreting complex Tsetlin Machines with a large number of clauses.
method Developed closed-form expressions for local and global interpretability of Tsetlin Machines.
result The expressions enable real-time feature importance assessment and data clustering.
Method screens similar capsule endoscopic images, reducing doctor workload and improving accuracy.
problem Time-consuming and high error rate in manual inspection of large numbers of similar capsule endoscopic images.
method Structural similarity analysis of visually salient areas and hierarchical clustering.
result 76% reduction in similar images, 100% lesion recall, 18-minute average play time.
Study uses random matrix test to find significant factors in cryptocurrency forecasts.
problem Determining the optimal number of factors in cryptocurrency forecast models.
method Applied a random matrix test to a forecast model of Reduced Rank Regression (RRR) on cryptocurrencies.
result Consistent results with visual inspection, minimal computational cost compared to cross-validation.
In this paper, we propose a novel generative model named Stacked Generative Adversarial Networks (SGAN), which is trained to invert the hierarchical representations of a bottom-up discriminative network. Our model consists of a top-down stack of GANs, each learned to generate lower-level representations conditioned on …
Deep learning improves defect classification in real-time surface inspection.
problem Real-time defect classification in manufacturing industry using limited datasets.
method Convolutional Neural Networks (CNNs) designed for speed and accuracy, neural data augmentation for class imbalance.
result 98.0% accuracy in binary defect classification with 22,000 labeled images.
A framework compares image representations based on local geometry.
problem Comparing image representations based on global structure overlooks local differences.
method Quantify local geometry using Fisher information matrix and optimize differentiation with principal distortions.
result Identifies differences in local sensitivities between models.
A new method extracts linguistic objects from text using CNNs.
problem Lack of interpretability in deep learning models for text.
method Weighted extension of Text Deconvolution Saliency (wTDS) measure.
result Extracts interpretable linguistic objects from text.
Deep learning detects corrosion in nuclear fuel canisters.
problem Ensuring safety and integrity of used nuclear fuel dry storage canisters.
method Residual neural networks (ResNets) for real-time corrosion detection of canister images.
result Deep learning approach accurately detects corrosion and classifies canisters as corroded or intact.
We improve GANs by enforcing reproducibility and using non-uniform sampling.
problem Overrepresentation of certain samples in GANs' marginal log-likelihood.
method Enforce reproducibility through matching empirical distribution to prior, use non-uniform sampling for mini-batch selection.
result Improved quality and variety in generated samples, validated on CIFAR10, Fashion MNIST, and CelebA.
It can be difficult to tell whether a trained generative model has learned to generate novel examples or has simply memorized a specific set of outputs. In published work, it is common to attempt to address this visually, for example by displaying a generated example and its nearest neighbor(s) in the training set (in,…
DivDis learns diverse hypotheses from underspecified data to improve robustness.
problem Learning from underspecified datasets leads to multiple equally viable solutions, causing out-of-distribution issues.
method DivDis framework: 1) learns diverse hypotheses using unlabeled test data, 2) selects one hypothesis with minimal additional supervision.
result DivDis finds robust features in image and natural language processing problems.
Feature engineering remains a major bottleneck when creating predictive systems from electronic medical records. At present, an important missing element is detecting predictive regular clinical motifs from irregular episodic records. We present Deepr (short for Deep record), a new end-to-end deep learning system that …
Many optimization algorithms converge to stationary points. When the underlying problem is nonconvex, they may get trapped at local minimizers and occasionally stagnate near saddle points. We propose the Run-and-Inspect Method, which adds an "inspect" phase to existing algorithms that helps escape from non-global stati…
New insights into how to inspect and learn from multi-stage processes and AI reasoning.
problem Understanding how to attribute outcomes to early stages in multi-stage operations and AI reasoning.
method Information-theoretic analysis and mathematical proofs of four key results.
result Uniform checkpoint spacing is minimax-optimal for inspection design under homogeneous signal attenuation.
Paper proposes machine learning models for more accurate road inspection.
problem Traditional road inspection systems have safety, energy, and cost issues.
method Hybrid machine learning models using surface deflection data from FWD tests.
result CMIS model outperforms other models with APRE=2.3303, AAPRE=11.6768, RMSE=12.0056, and SD=0.0210.
Reliably detecting anomalies in a given set of images is a task of high practical relevance for visual quality inspection, surveillance, or medical image analysis. Autoencoder neural networks learn to reconstruct normal images, and hence can classify those images as anomalies, where the reconstruction error exceeds som…
ERICA assesses replicability of cluster analysis results.
problem Lack of quantitative scrutiny for clustering results.
method ERICA: a framework to assess replicability of cluster analysis.
result Clusters are found to be replicable in synthetic data but not in real-world datasets.
Enhances privacy in data annotation and inspection with synthetic data.
problem Improving privacy in machine learning tasks like data annotation and inspection.
method Employing Bayesian differential privacy to generate higher-fidelity synthetic data.
result Produces higher-fidelity samples, detecting more subtle data errors and biases.
The predictive power of neural networks often costs model interpretability. Several techniques have been developed for explaining model outputs in terms of input features; however, it is difficult to translate such interpretations into actionable insight. Here, we propose a framework to analyze predictions in terms of …
Generative models help debug data issues in private, decentralized datasets.
problem Manual data inspection is impractical for private, decentralized datasets.
method Differentially private federated generative models.
result Generative models can effectively debug data issues in private datasets.