Convolutional Neural Networks are widely used in various machine learning domains. In image processing, the features can be obtained by applying 2D convolution to all spatial dimensions of the input. However, in the audio case, frequency domain input like Mel-Spectrogram has different and unique characteristics in the frequency dimension. Thus, there is a need for a method that allows the 2D convolution layer to handle the frequency dimension differently. In this work, we introduce SubSpectral Normalization (SSN), which splits the input frequency dimension into several groups (sub-bands) and performs a different normalization for each group. SSN also includes an affine transformation that can be applied to each group. Our method removes the inter-frequency deflection while the network learns a frequency-aware characteristic. In the experiments with audio data, we observed that SSN can efficiently improve the network's performance.

Class labels used for machine learning are relatable to each other, with certain class labels being more similar to each other than others (e.g. images of cats and dogs are more similar to each other than those of cats and cars). Such similarity among classes is often the cause of poor model performance due to the models confusing between them. Current labeling techniques fail to explicitly capture such similarity information. In this paper, we instead exploit the similarity between classes by capturing the similarity information with our novel confidence labels. Confidence labels are probabilistic labels denoting the likelihood of similarity, or confusability, between the classes. Often even after models are trained to differentiate between classes in the feature space, the similar classes' latent space still remains clustered. We view this type of clustering as valuable information and exploit it with our novel projective loss functions. Our projective loss functions are designed to work with confidence labels with an ability to relax the loss penalty for errors that confuse similar classes. We use our approach to train neural networks with noisy labels, as we believe noisy labels are partly a result of confusability arising from class similarity. We show improved performance compared to the use of standard loss functions. We conduct a detailed analysis using the CIFAR-10 dataset and show our proposed methods' applicability to larger datasets, such as ImageNet and Food-101N.

In this work, we test the performance of two bidirectional transformer-based language models, BERT and SpanBERT, on predicting directionality in causal pairs in the textual content. Our preliminary results show that predicting direction for inter-sentence and implicit causal relations is more challenging. And, SpanBERT performs better than BERT on causal samples with longer span length. We also introduce CREST which is a framework for unifying a collection of scattered datasets of causal relations.

Abstract--Stack Overflow is one of the most popular Programming Community-based Question Answering (PCQA) websites that has attracted more and more users in recent years. When users raise or inquire questions in Stack Overflow, providing related questions can help them solve problems. Although there are many approaches based on deep learning that can automatically predict the relatedness between questions, those approaches are limited since interaction information between two questions may be lost. In this paper, we adopt the deep learning technique, propose an Attention-based Sentence pair Interaction Model (ASIM) to predict the relatedness between questions on Stack Overflow automatically. We adopt the attention mechanism to capture the semantic interaction information between the questions. Besides, we have pre-trained and released word embeddings specific to the software engineering domain for this task, which may also help other related tasks. The experiment results demonstrate that ASIM has made significant improvement over the baseline approaches in Precision, Recall, and Micro-F1 evaluation metrics, achieving state-of-the-art performance in this task. Our model also performs well in the duplicate question detection task of AskUbuntu, which is a similar but different task, proving its generalization and robustness. The continuous progress of the information industry has caused more and more people to engage in software development. Therefore, Programming Community-based Question questions related to the question raised or being inquired by Answering (PCQA) websites have attracted a large number users can effectively help them solve problems.

In this work, we propose a novel deep learning framework that can generate a vivid dance from a whole piece of music. In contrast to previous works that define the problem as generation of frames of motion state parameters, we formulate the task as a prediction of motion curves between key poses, which is inspired by the animation industry practice. The proposed framework, named DanceNet3D, first generates key poses on beats of the given music and then predicts the in-between motion curves. DanceNet3D adopts the encoder-decoder architecture and the adversarial schemes for training. The decoders in DanceNet3D are constructed on MoTrans, a transformer tailored for motion generation. In MoTrans we introduce the kinematic correlation by the Kinematic Chain Networks, and we also propose the Learned Local Attention module to take the temporal local correlation of human motion into consideration. Furthermore, we propose PhantomDance, the first large-scale dance dataset produced by professional animatiors, with accurate synchronization with music. Extensive experiments demonstrate that the proposed approach can generate fluent, elegant, performative and beat-synchronized 3D dances, which significantly surpasses previous works quantitatively and qualitatively.

Robotic manipulation of unknown objects is an important field of research. Practical applications occur in many real-world settings where robots need to interact with an unknown environment. We tackle the problem of reactive grasping by proposing a method for unknown object tracking, grasp point sampling and dynamic trajectory planning. Our object tracking method combines Siamese Networks with an Iterative Closest Point approach for pointcloud registration into a method for 6-DoF unknown object tracking. The method does not require further training and is robust to noise and occlusion. We propose a robotic manipulation system, which is able to grasp a wide variety of formerly unseen objects and is robust against object perturbations and inferior grasping points.

The unification of statistical (data-driven) and symbolic (knowledge-driven) methods is widely recognised as one of the key challenges of modern AI. Recent years have seen large number of publications on such hybrid neuro-symbolic AI systems. That rapidly growing literature is highly diverse and mostly empirical, and is lacking a unifying view of the large variety of these hybrid systems. In this paper we analyse a large body of recent literature and we propose a set of modular design patterns for such hybrid, neuro-symbolic systems. We are able to describe the architecture of a very large number of hybrid systems by composing only a small set of elementary patterns as building blocks. The main contributions of this paper are: 1) a taxonomically organised vocabulary to describe both processes and data structures used in hybrid systems; 2) a set of 15+ design patterns for hybrid AI systems, organised in a set of elementary patterns and a set of compositional patterns; 3) an application of these design patterns in two realistic use-cases for hybrid AI systems. Our patterns reveal similarities between systems that were not recognised until now. Finally, our design patterns extend and refine Kautz' earlier attempt at categorising neuro-symbolic architectures.

In recent years, GANs (generative adversarial networks) have been all the rage in the field of deep-learning generative models, leaving VAEs in relative obscurity. But there's much to gain from a solid footing in variational autoencoders, which tackle similar challenges but use a different architectural foundation. If you were looking for an engaging, accessible way to learn more about VAEs, Joseph and Baptiste Rocca's introduction hits the spot. They define terms, walk us through the various elements that make up VAEs and how they relate to each other, and add beautiful illustrations for all the visual learners out there.

CNNs or convolutional neural nets are a type of deep learning algorithm that does really well at learning images. That's because they can learn patterns that are translation invariant and have spatial hierarchies (F. That means if If the CNN learns the dog in the left corner of the image above, then it can identify the dog in the other two pictures that have been moved around (translation invariance). If the CNN learns the dog from the left corner of the image above, it will recognize pieces of the original image in the other two pictures because it has learned what the edges of the her eye with heterochromia looks like, her wolf-like snout and the shape of her stylish headphones (spatial hierarchies). These properties make CNNs formidable learners for images because the real world doesn't always look exactly like the training data.

Quarantine has made a lot of us rusty in the flirting department, but your weak game is nothing compared to the cringeworthy pickup lines generated by an artificial intelligence. Research scientist Janelle Shane recently tested how the world's largest language modeling software, called GPT-3, would handle something as nuanced as flirting. Shane tasked different versions of the algorithm to come up with their best pickup lines, but the results would hardly make anyone weak in the knees. 'I will briefly summarize the plot of Back to the Future II for you,' offered another. An algorithm tasked with coming up with sexy pick-up lines had less than romantic results: 'I will briefly summarize the plot of'Back to the Future II' for you,' read one ice-breaker Shane, an optics research scientist at Boulder Nonlinear Systems, likes to tinker with what she calls'AI weirdness,' experiments in machine-learning that often have hilarious results.