VCML learns concepts and metaconcepts from images and questions.
problem Learning concepts and metaconcepts from visual data.
method Bidirectional connection between visual concepts and metaconcepts.
result VCML can generalize from limited data and noisy inputs.
The paper explains knowledge distillation by analyzing visual concepts in DNNs.
problem Understanding how knowledge distillation affects the learning of visual concepts in deep neural networks.
method The paper proposes three hypotheses and designs mathematical metrics to evaluate feature representations of DNNs.
result The hypotheses were verified through experiments on various DNNs.
More than 50 years ago Bongard introduced 100 visual concept learning problems as a testbed for intelligent vision systems. These problems are now known as Bongard problems. Although they are well known in the cognitive science and AI communities only moderate progress has been made towards building systems that can so…
Study explores how to explain deep neural networks using interactive naming.
problem Explaining deep neural networks' decisions in human-understandable terms.
method Developed an interactive naming interface for human annotators to cluster activation maps into visual concepts.
result Significant agreement among annotators about visual concepts, many activation maps have recognizable concepts.
Visual analytics tool detects and corrects concept drift in data streams.
problem Concept drift causes inaccurate predictions in evolving data.
method DriftVis combines drift detection and visualization.
result Visual analytics supports detection, examination, and correction of concept drift.
The seemingly infinite diversity of the natural world arises from a relatively small set of coherent rules, such as the laws of physics or chemistry. We conjecture that these rules give rise to regularities that can be discovered through primarily unsupervised experiences and represented as abstract concepts. If such r…
Algorithm transfers visual concepts to answer out-of-vocabulary questions.
problem Leveraging off-the-shelf visual and linguistic data for out-of-vocabulary answers in visual question answering.
method Unsupervised task discovery for learning task conditional visual classifier, then transferring to visual question answering models.
result Algorithm generalizes to out-of-vocabulary answers successfully.
Concept Hierarchies and Formal Concept Analysis are theoretically well grounded and largely experimented methods. They rely on line diagrams called Galois lattices for visualizing and analysing object-attribute sets. Galois lattices are visually seducing and conceptually rich for experts. However they present important…
The task of associating images and videos with a natural language description has attracted a great amount of attention recently. Rapid progress has been made in terms of both developing novel algorithms and releasing new datasets. Indeed, the state-of-the-art results on some of the standard datasets have been pushed i…
A framework visualizes embedding spaces of neural survival analysis models using anchor directions.
problem Visualizing complex embeddings in neural survival analysis models.
method Estimating anchor directions through clustering or user-supplied concepts, revealing relationships with raw inputs and survival times.
result Visualization strategies reveal how anchor directions relate to raw clinical features and survival time distributions.
Develops ACE to automatically identify meaningful concepts from neural network predictions.
problem Challenges in interpreting feature importance scores for machine learning models.
method Proposes concept-based explanation principles and develops ACE algorithm to extract visual concepts.
result Demonstrates ACE discovers human-meaningful, coherent concepts for neural network predictions.
People can learn complex visual concepts from just a few examples.
problem Understanding how people learn and categorize visual concepts from limited data.
method Bayesian program learning model that searches for the best explanation of observations.
result People's judgments are broadly consistent with a Bayesian program learning model, indicating they can learn rich algorithmic abstractions from sparse input data.
A neural network visualizes data structure and concepts.
problem Data visualization and concept understanding.
method Mixing autoencoder and classifier for multi-perspective visualization.
result The network produces different topological maps based on training as autoencoder or classifier.
New method for visualizing high-level concepts in generative models.
problem Challenges in evaluating and visualizing concepts in generative models.
method Introduces a method to compute concept saliency maps for latent representations of known or novel high-level concepts.
result Concept saliency maps highlight input features important for high-level concepts.
Enhances visual explanations with logical rules for complex concepts.
problem Lack of explanatory power for deep learning models, especially for relational concepts.
method Combining LIME for highlighting and Aleph for logic rules.
result Generated relational rules can be explicitly linked to input images.
Visualizes board connections for socially responsible investing insights.
problem Understanding corporate governance and sustainability through board connections.
method Data Visualization tool to reveal connections between Directors and Executives.
result Strength of tool in investigating corporate governance and sustainability.
NegToMe uses images to guide text-based models away from unwanted visual elements.
problem Insufficient text-based adversarial guidance for complex visual concepts.
method Negative token merging (NegToMe) using visual features from reference images.
result Significantly enhances output diversity and reduces visual similarity to copyrighted content.
Automated discovery of early visual concepts from raw image data is a major open challenge in AI research. Addressing this problem, we propose an unsupervised approach for learning disentangled representations of the underlying factors of variation. We draw inspiration from neuroscience, and show how this can be achiev…
Measuring the distance between concepts is an important field of study of Natural Language Processing, as it can be used to improve tasks related to the interpretation of those same concepts. WordNet, which includes a wide variety of concepts associated with words (i.e., synsets), is often used as a source for computin…
Proposes a method to measure model parameter similarity for visual tasks.
problem Estimating relations between different visual tasks.
method LPS method using a second-order neural network to align model parameters and learn second-order similarity.
result Extensive experiments validate the effectiveness of the proposed method.
New method evaluates text-to-image synthesis for realism, variety, and semantic accuracy.
problem Lack of metrics revealing semantic accuracy in text-to-image synthesis.
method Uses Inception network representations and t-SNE visualization for semantic evaluation.
result Classification accuracy of generated images to real images' visual concepts correlates with semantic accuracy.
It is easy for people to imagine what a man with pink hair looks like, even if they have never seen such a person before. We call the ability to create images of novel semantic concepts visually grounded imagination. In this paper, we show how we can modify variational auto-encoders to perform this task. Our method use…
Automated generation of medical reports from chest x-rays using expert annotations.
problem Generating long, unstructured text from medical images with context and consistency.
method First learn visually-informative medical concepts from raw reports, then use these concepts to auto-generate structured reports from images.
result Validation on OpenI dataset shows auto-generated reports are consistent with manual annotations.
Study uses crochet to visualize non-Euclidean geometry.
problem Understanding non-Euclidean surfaces through physical models.
method Parametrization of crochet models to represent Lobachevskian surface.
result Crochet models reflect non-Euclidean geometry characteristics.
Model learns multiple tasks using visual and textual representations.
problem Training visual navigation agents for multiple tasks.
method Dual-Attention unit for task-invariant alignment of visual and textual representations.
result Model outperforms baselines on semantic goal navigation and embodied question answering.
We provide complete source code for a front-end GUI and its back-end counterpart for a stock market visualization tool. It is built based on the "functional visualization" concept we discuss, whereby functionality is not sacrificed for fancy graphics. The GUI, among other things, displays a color-coded signal (computed…
ViCE uses superpixels to enhance self-supervised learning for better dense visual embeddings.
problem Lack of high-resolution feature maps from self-supervised models.
method Superpixels for dense representation learning, contrasting over regions.
result Improves unsupervised semantic segmentation on benchmarks like Cityscapes and COCO.
ARNe model excels in abstract visual reasoning tasks.
problem Abstract visual reasoning using attention mechanisms.
method Hybrid network architecture combining self-attention and relational reasoning.
result ARNe model surpasses WReN model by 11.28 ppt on PGM datasets.
Deep neural networks are complex and opaque. As they enter application in a variety of important and safety critical domains, users seek methods to explain their output predictions. We develop an approach to explaining deep neural networks by constructing causal models on salient concepts contained in a CNN. We develop…
Convolutional neural networks (CNNs) are one of the driving forces for the advancement of computer vision. Despite their promising performances on many tasks, CNNs still face major obstacles on the road to achieving ideal machine intelligence. One is that CNNs are complex and hard to interpret. Another is that standard…
Energy-based models can generate complex images by combining simpler concepts.
problem Generating natural images that satisfy complex logical combinations of concepts.
method Energy-based models combine probability distributions of simpler concepts to generate compositions.
result Energy-based models can generate images that satisfy conjunctions, disjunctions, and negations of concepts.
We propose Deep Feature Factorization (DFF), a method capable of localizing similar semantic concepts within an image or a set of images. We use DFF to gain insight into a deep convolutional neural network's learned features, where we detect hierarchical cluster structures in feature space. This is visualized as heat m…
I propose and briefly define the concept of Urban Isobenefit Lines by using functions as easy as efficient, whose results can offer a rich tool to use into spatial equilibrium analysis involving cities. They are line joining urban points with equal level of positional advantage from city amenities. The results which on…
Enhances group convolutional networks with attention to learn meaningful relationships.
problem Lack of explicit means to learn meaningful relationships among symmetry patterns.
method Introduces attentive group equivariant convolutions, applying attention during convolution.
result Consistently outperforms conventional group convolutional networks on benchmark datasets.
The paper explores methods to explain brain tumor segmentation models and extract visualizations.
problem Improving interpretability of deep learning models for brain tumor segmentation.
method Exploring techniques to explain brain tumor segmentation models and extract visualizations.
result Brain tumor segmentation networks learn human-understandable disentangled concepts and use a top-down approach.
Improves generative Visual Dialog by asking diverse questions.
problem Generative Visual Dialog models degrade after a few rounds of interaction.
method Introduce a simple auxiliary objective to incentivize Qbot to ask diverse questions.
result Better dialog diversity, consistency, fluency, and detail with improved image relevance.
A deep model generates diverse graph layouts.
problem Finding a good graph layout is time-consuming and haphazard.
method Deep generative model learns latent space for diverse layouts.
result Model learns and generalizes graph layout concepts.
In an effort to understand the meaning of the intermediate representations captured by deep networks, recent papers have tried to associate specific semantic concepts to individual neural network filter responses, where interesting correlations are often found, largely by focusing on extremal filter responses. In this …
Analyzes GAN units for better understanding and improvement.
problem Lack of understanding GAN internal representations and artifacts.
method Interpretable units identification, causal effect quantification, contextual relationship examination.
result Visualizes and understands GANs at various levels, enabling new insights and improvements.
Variational methods for revealing visual concepts learned by convolutional neural networks have gained significant attention during the last years. Being based on noisy gradients obtained via back-propagation such methods require the application of regularization strategies. We present a mathematical framework unifying…
CA connects visual concepts to neural network representations.
problem Global association of human interpretable concepts with neural network internals.
method Established a taxonomy for CA methods and reviewed state-of-the-art research.
result Unified understanding of CA methods and datasets.
Most research on the interpretability of machine learning systems focuses on the development of a more rigorous notion of interpretability. I suggest that a better understanding of the deficiencies of the intuitive notion of interpretability is needed as well. I show that visualization enables but also impedes intuitiv…
Low dimensional embeddings that capture the main variations of interest in collections of data are important for many applications. One way to construct these embeddings is to acquire estimates of similarity from the crowd. However, similarity is a multi-dimensional concept that varies from individual to individual. Ex…
Hybrid model learns novel handwritten characters better than neural or symbolic models alone.
problem Generating novel yet structured concepts.
method Neuro-symbolic model combining neural networks and probabilistic programs.
result Hybrid model outperforms alternative models in learning and generalizing novel handwritten characters.
ivhd tool efficiently visualizes large high-dimensional data.
problem Interactive visualization of large high-dimensional datasets.
method Minimal memory and computational time requirements for nearest neighbor visualization.
result Significantly reduces computational and memory loads.
MXGNet tackles visual reasoning tasks using graph neural networks.
problem Abstract reasoning, especially in the visual domain, is challenging for AI.
method Combines object-level representations, graph neural networks, and multiplex graphs.
result Achieves state-of-the-art accuracy on Euler Diagram Syllogisms and outperforms state-of-the-art models on RPM datasets.
SketchEmbedNet learns image representations from sketches, useful for few-shot learning.
problem Learning image representations from sketches for few-shot learning.
method Training a model to produce sketches of images, focusing on informative embeddings.
result Model produces informative embeddings of novel images, classes, and datasets.
Proposes a fuzzy rule-based method for data visualization.
problem Preserving neighborhood relationships and handling non-linear manifolds in data visualization.
method Uses a first-order Takagi-Sugeno model with clusters and Geodesic c-means clustering for rule generation and parameter estimation.
result Behaves desirably and performs better than or comparable to other methods.