Deep Learning (DL) has shown great promise in the unsupervised task of clustering. That said, while in classical (i.e., non-deep) clustering the benefits of the nonparametric approach are well known, most deep-clustering methods are parametric: namely, they require a predefined and fixed number of clusters, denoted by K. When K is unknown, however, using model-selection criteria to choose its optimal value might become computationally expensive, especially in DL as the training process would have to be repeated numerous times. In this work, we bridge this gap by introducing an effective deep-clustering method that does not require knowing the value of K as it infers it during the learning. Using a split/merge framework, a dynamic architecture that adapts to the changing K, and a novel loss, our proposed method outperforms existing nonparametric methods (both classical and deep ones). While the very few existing deep nonparametric methods lack scalability, we demonstrate ours by being the first to report the performance of such a method on ImageNet. We also demonstrate the importance of inferring K by showing how methods that fix it deteriorate in performance when their assumed K value gets further from the ground-truth one, especially on imbalanced datasets. Our code is available at https://github.com/BGU-CS-VIL/DeepDPM.

Abstract. Brain computer interface (BCI) systems are developed in the biomedical engineering fields to increase the quality of life among patients with paralysis and neurological conditions. The development of a six class BCI controller to operate a semi-autonomous mobile robotic arm is presented. The controller uses the following mental tasks: imagined left/right hand squeeze, imagined left/right foot tap, rest, and a physical jaw clench. To design a controller, the locations of active electrodes are verified, and an appropriate machine learning algorithm is determined. Three subjects, ages ranging between 22 and 27, participated in five sessions of motor imagery experiments to record their brainwaves. These recordings were analyzed using event related potential (ERP) plots and topographical maps to determine active electrodes. bcilab was used to train two, three, five, and six class BCI controllers using linear discriminant analysis (LDA) and relevance vector machine (RVM) machine learning methods. The subjects' data were used to compare the two-method's performance in terms of error rate percentage. While a two class BCI controller showed the same accuracy for both methods, the three and five class BCI controllers showed the RVM approach having a higher accuracy than the LDA approach. For the five-class controller, error rate percentage was 33.3% for LDA and 29.2% for RVM. The six class BCI controller error rate percentage for both LDA and RVM was 34.5%. While the percentage values are the same, RVM was chosen as the desired machine learning algorithm based on the trend seen in the three and five class controller performances.

Data science offers all the opportunities of the hottest profession to the current tech-savvy generation, who are highly interested to work in data science than traditional engineering careers. Data analytics courses are in huge demand among the courses for data professionals. Students can access multiple online data science courses on multiple educational platforms having collaborations with reputed educational institutes. Courses on data science are providing a sufficient and deep understanding of all key concepts and hands-on experience with real-life projects to candidates. Let's explore some of the top trending data science courses in 2022 to take up to become a competent data science professional.

In this post, I will provide an overview of generalized additive models (GAMs) and their desirable features. Predictive accuracy has long been an important goal of machine learning. But model interpretability has received more attention in recent years. Stakeholders, such as executives, regulators, and domain experts, often want to understand how and why a model makes its predictions before they trust it enough to use it in practice. However, when you train a machine learning model, you typically face a tradeoff between accuracy and interpretability.

One particular problem arises at this stage. How do we know what model to choose as the meta-model? Unfortunately, there has't been any research into this area and the selection of a meta-model is more of an art than a science. In most of the papers discussing stacked models, the meta-model used is often just a simple model such as Linear Regression for regression tasks and Logistic Regression for classification tasks. One reason why more complex meta-models are often not chosen is because there is a much higher chance that the meta-model may overfit to the predictions from the base models.

Brain-computer interfaces (BCIs), transform neural signals in the brain into in-structions to control external devices. However, obtaining sufficient training data is difficult as well as limited. With the advent of advanced machine learning methods, the capability of brain-computer interfaces has been enhanced like never before, however, these methods require a large amount of data for training and thus require data augmentation of the limited data available. Here, we use spiking neural networks (SNN) as data generators. It is touted as the next-generation neu-ral network and is considered as one of the algorithms oriented to general artifi-cial intelligence because it borrows the neural information processing from bio-logical neurons. We use the SNN to generate neural spike information that is bio-interpretable and conforms to the intrinsic patterns in the original neural data. Ex-periments show that the model can directly synthesize new spike trains, which in turn improves the generalization ability of the BCI decoder. Both the input and output of the spiking neural model are spike information, which is a brain-inspired intelligence approach that can be better integrated with BCI in the future.

Designing effective rehabilitation strategies for upper extremities, particularly hands and fingers, warrants the need for a computational model of human motor learning. The presence of large degrees of freedom (DoFs) available in these systems makes it difficult to balance the trade-off between learning the full dexterity and accomplishing manipulation goals. The motor learning literature argues that humans use motor synergies to reduce the dimension of control space. Using the low-dimensional space spanned by these synergies, we develop a computational model based on the internal model theory of motor control. We analyze the proposed model in terms of its convergence properties and fit it to the data collected from human experiments. We compare the performance of the fitted model to the experimental data and show that it captures human motor learning behavior well.

Probabilistic word embedding models have emerged as a way to model textual data in scientific applications [Rudolph We introduce probabilistic embeddings using et al., 2016, Bamler and Mandt, 2017]. The advantages include Laplacian priors (PELP). The proposed model enables flexible inclusion of prior knowledge, explicit handling incorporating graph side-information into of uncertainty, straightforward estimation, and usefulness static word embeddings. We theoretically show in decision-making [Ghahramani, 2015]. More versatile that the model unifies several previously proposed and flexible word embedding methods allow for increasingly embedding methods under one umbrella.

Deep neural networks (DNNs) have found their way into many applications with potential impact on the safety, security, and fairness of human-machine-systems. Such require basic understanding and sufficient trust by the users. This motivated the research field of explainable artificial intelligence (XAI), i.e. finding methods for opening the "black-boxes" DNNs represent. For the computer vision domain in specific, practical assessment of DNNs requires a globally valid association of human interpretable concepts with internals of the model. The research field of concept (embedding) analysis (CA) tackles this problem: CA aims to find global, assessable associations of humanly interpretable semantic concepts (e.g., eye, bearded) with internal representations of a DNN. This work establishes a general definition of CA and a taxonomy for CA methods, uniting several ideas from literature. That allows to easily position and compare CA approaches. Guided by the defined notions, the current state-of-the-art research regarding CA methods and interesting applications are reviewed. More than thirty relevant methods are discussed, compared, and categorized. Finally, for practitioners, a survey of fifteen datasets is provided that have been used for supervised concept analysis. Open challenges and research directions are pointed out at the end.

Traditional statistical learning theory relies on the assumption that data are identically and independently generated from a given distribution (i.i.d.). The independently distributed assumption, on the other hand, fails to hold in many real applications. In this survey, we consider learning settings in which examples are dependent and their dependence relationship can be characterized by a graph. We collect various graph-dependent concentration bounds, which are then used to derive Rademacher and stability generalization bounds for learning from graph-dependent data. We illustrate this paradigm with three learning tasks and provide some research directions for future work. To the best of our knowledge, this is the first survey on this subject.