We propose a connectionist-inspired kernel machine model with three key advantages over traditional kernel machines. First, it is capable of learning distributed and hierarchical representations. Second, its performance is highly robust to the choice of kernel function. Third, the solution space is not limited to the span of images of training data in reproducing kernel Hilbert space (RKHS). Together with the architecture, we propose a greedy learning algorithm that allows the proposed multilayer network to be trained layer-wise without backpropagation by optimizing the geometric properties of images in RKHS. With a single fixed generic kernel for each layer and two layers in total, our model compares favorably with state-of-the-art multiple kernel learning algorithms using significantly more kernels and popular deep architectures on widely used classification benchmarks.

In spoken dialogue systems, we aim to deploy artificial intelligence to build automated dialogue agents that can converse with humans. A part of this effort is the policy optimisation task, which attempts to find a policy describing how to respond to humans, in the form of a function taking the current state of the dialogue and returning the response of the system. In this paper, we investigate deep reinforcement learning approaches to solve this problem. Particular attention is given to actor-critic methods, off-policy reinforcement learning with experience replay, and various methods aimed at reducing the bias and variance of estimators. When combined, these methods result in the previously proposed ACER algorithm that gave competitive results in gaming environments. These environments however are fully observable and have a relatively small action set so in this paper we examine the application of ACER to dialogue policy optimisation. We show that this method beats the current state-of-the-art in deep learning approaches for spoken dialogue systems. This not only leads to a more sample efficient algorithm that can train faster, but also allows us to apply the algorithm in more difficult environments than before. We thus experiment with learning in a very large action space, which has two orders of magnitude more actions than previously considered. We find that ACER trains significantly faster than the current state-of-the-art.

It has been widely observed that many activation functions and pooling methods of neural network models have (positive-) rescaling-invariant property, including ReLU, PReLU, max-pooling, and average pooling, which makes fully-connected neural networks (FNNs) and convolutional neural networks (CNNs) invariant to (positive) rescaling operation across layers. This may cause unneglectable problems with their optimization: (1) different NN models could be equivalent, but their gradients can be very different from each other; (2) it can be proven that the loss functions may have many spurious critical points in the redundant weight space. To tackle these problems, in this paper, we first characterize the rescaling-invariant properties of NN models using equivalent classes and prove that the dimension of the equivalent class space is significantly smaller than the dimension of the original weight space. Then we represent the loss function in the compact equivalent class space and develop novel algorithms that conduct optimization of the NN models directly in the equivalent class space. We call these algorithms Equivalent Class Optimization (abbreviated as EC-Opt) algorithms. Moreover, we design efficient tricks to compute the gradients in the equivalent class, which almost have no extra computational complexity as compared to standard back-propagation (BP). We conducted experimental study to demonstrate the effectiveness of our proposed new optimization algorithms. In particular, we show that by using the idea of EC-Opt, we can significantly improve the accuracy of the learned model (for both FNN and CNN), as compared to using conventional stochastic gradient descent algorithms.

We propose Efficient Neural Architecture Search (ENAS), a fast and inexpensive approach for automatic model design. In ENAS, a controller learns to discover neural network architectures by searching for an optimal subgraph within a large computational graph. The controller is trained with policy gradient to select a subgraph that maximizes the expected reward on the validation set. Meanwhile the model corresponding to the selected subgraph is trained to minimize a canonical cross entropy loss. Thanks to parameter sharing between child models, ENAS is fast: it delivers strong empirical performances using much fewer GPU-hours than all existing automatic model design approaches, and notably, 1000x less expensive than standard Neural Architecture Search. On the Penn Treebank dataset, ENAS discovers a novel architecture that achieves a test perplexity of 55.8, establishing a new state-of-the-art among all methods without post-training processing. On the CIFAR-10 dataset, ENAS designs novel architectures that achieve a test error of 2.89%, which is on par with NASNet (Zoph et al., 2018), whose test error is 2.65%.

Spectral graph convolutional neural networks (CNNs) require approximation to the convolution to alleviate the computational complexity, resulting in performance loss. This paper proposes the topology adaptive graph convolutional network (TAGCN), a novel graph convolutional network defined in the vertex domain. We provide a systematic way to design a set of fixed-size learnable filters to perform convolutions on graphs. The topologies of these filters are adaptive to the topology of the graph when they scan the graph to perform convolution. The TAGCN not only inherits the properties of convolutions in CNN for grid-structured data, but it is also consistent with convolution as defined in graph signal processing. Since no approximation to the convolution is needed, TAGCN exhibits better performance than existing spectral CNNs on a number of data sets and is also computationally simpler than other recent methods.

Imagine this scenario: In a clinical trial investigating the toxicity of a new chemotherapy drug to treat breast cancer, some patients drop out of the trial before completion for various reasons, hence we do not have the data for final outcome on the dropped out patients. What if the patients who drop out of the trial before completion are the ones who experienced toxicity and are unwilling to continue the treatment, this reason however is not recorded in the database and the patients are marked as "lost to followup". If the investigators were to analyze the data using conventional methods where loss to followup is ignored and not properly accounted for, they will estimate the toxicity to be far less than what it really is. These results can lead to adapting a drug, that is otherwise unsafe. Similarly if patients who are feeling better dropout of the trial before completion, the estimates of toxicity would be far greater than the real value, leading to rejection of a potential lifesaver drug.

Bayesian Optimisation (BO) refers to a class of methods for global optimisation of a function $f$ which is only accessible via point evaluations. It is typically used in settings where $f$ is expensive to evaluate. A common use case for BO in machine learning is model selection, where it is not possible to analytically model the generalisation performance of a statistical model, and we resort to noisy and expensive training and validation procedures to choose the best model. Conventional BO methods have focused on Euclidean and categorical domains, which, in the context of model selection, only permits tuning scalar hyper-parameters of machine learning algorithms. However, with the surge of interest in deep learning, there is an increasing demand to tune neural network \emph{architectures}. In this work, we develop NASBOT, a Gaussian process based BO framework for neural architecture search. To accomplish this, we develop a distance metric in the space of neural network architectures which can be computed efficiently via an optimal transport program. This distance might be of independent interest to the deep learning community as it may find applications outside of BO. We demonstrate that NASBOT outperforms other alternatives for architecture search in several cross validation based model selection tasks on multi-layer perceptrons and convolutional neural networks.

I am a developer, and I've (as you've) heard a lot about machine learning and neural networks ... and whenever I decide to take my first step, I find myself dealing with something like: They are also known as shift invariant or space invariant artificial neural networks (SIANN), based on their shared-weights architecture and translation invariance characteristics. This kind of explanations pushed me a while from moving on, until I read some articles and saw some talks (like this one), confirming that Machine learning should be more fun. Before continuing, please keep in mind that I'm not an expert,I'm far away from that, I'm just counting my very first steps and wanted to share my experience throughout this article about building and using CNN using Deeplearning4j. As a human, we intuitively know that pictures have a hierarchy or conceptual structure. Most importantly, we recognize the idea of a bird no matter what position the bird take.

On GitHub, developers are spending more time on cross-platform development, deep learning, and acquring new coding skills, based on the popular code-sharig site's analysis of activity on GiHub in 2017. Google's Angular JavaScript framework and TensorFlow machine learning library have experienced upticks in participation. Projects for learning, such as GitHub's Coding Interview University, also have done well. For multiplatform and web development, the Angular CLI project, which provides a command-line interface for Angular applications, more than doubled its number of contributors from 2016 to 2017. Interest in the Angular framework itself also increased, as did participation in Facebook's React JavaScript UI library and GitHub's own Electron framework.

Convolutional Neural Network(CNN) is a type of neural network especially useful for image classification tasks. I applied CNN on thousands of Simpsons images training the classifier to recognise 10 characters from the TV show with an accuracy of more than 90 percent. The dataset contains images for more than 20 Simpson characters. I picked the ones with a minimum of 1000 sample images and chose 10 characters to train my model. All chosen characters have thousand sample images.