Downloads & Free Reading Options - Results

Speaker intent detection and semantic slot filling are two critical tasks in spoken language understanding (SLU) for dialogue systems. In this paper, we describe a recurrent neural network (RNN) model that jointly performs intent detection, slot filling, and language modeling. The neural network model keeps updating the intent estimation as word in the transcribed utterance arrives and uses it as contextual features in the joint model. Evaluation of the language model and online SLU model is made on the ATIS benchmarking data set. On language modeling task, our joint model achieves 11.8% relative reduction on perplexity comparing to the independent training language model. On SLU tasks, our joint model outperforms the independent task training model by 22.3% on intent detection error rate, with slight degradation on slot filling F1 score. The joint model also shows advantageous performance in the realistic ASR settings with noisy speech input.

Increasing the capacity of recurrent neural networks (RNN) usually involves augmenting the size of the hidden layer, resulting in a significant increase of computational cost. An alternative is the recurrent neural tensor network (RNTN), which increases capacity by employing distinct hidden layer weights for each vocabulary word. However, memory usage scales linearly with vocabulary size, which can reach millions for word-level language models. In this paper, we introduce restricted recurrent neural tensor networks (r-RNTN) which reserve distinct hidden layer weights for frequent vocabulary words while sharing a single set of weights for infrequent words. Perplexity evaluations show that r-RNTNs improve language model performance over standard RNNs using only a small fraction of the parameters of unrestricted RNTNs.

In this episode, we discuss the problem of predicting the future from not only recent events but also from the distant past using Recurrent Neural Networks (RNNs). A example RNN is described which learns to label images with simple sentences. A learning machine capable of generating even simple descriptions of images such as these could be used to help the blind interpret images, provide assistance to children and adults in language acquisition, support internet search of content in images, and enhance search engine optimization websites containing unlabeled images. Both tutorial notes and advanced implementational notes for RNNs can be found in the show notes at: www.learningmachines101.com .

This study assesses the extent to which machine learning techniques can be used to predict question quality. An algorithm based on textual complexity indices was previously developed to assess question quality to provide feedback on questions generated by students within iSTART (an intelligent tutoring system that teaches reading strategies). In this study, 4,575 questions were coded by human raters based on their corresponding depth, classifying questions into four categories: 1-very shallow to 4-very deep. Here we propose a novel approach to assessing question quality within this dataset based on Recurrent Neural Networks (RNNs) and word embeddings. The experiments evaluated multiple RNN architectures using GRU, BiGRU and LSTM cell types of different sizes, and different word embeddings (i.e., FastText and Glove). The most precise model achieved a classification accuracy of 81.22%, which surpasses the previous prediction results using lexical sophistication complexity indices (accuracy = 41.6%). These results are promising and have implications for the future development of automated assessment tools within computer-based learning environments. [This is a paper in: Penstein Rosé, C., et al. "Artificial Intelligence in Education." AIED 2018. Lecture Notes in Computer Science(), vol 10947. Springer, Cham.]

A stationary state replica analysis for a dual neural network model that interpolates between a fully recurrent symmetric attractor network and a strictly feed-forward layered network, studied by Coolen and Viana, is extended in this work to account for finite dilution of the recurrent Hebbian interactions between binary Ising units within each layer. Gradual dilution is found to suppress part of the phase transitions that arise from the competition between recurrent and feed-forward operation modes of the network. Despite that, a long chain of layers still exhibits a relatively good performance under finite dilution for a balanced ratio between inter-layer and intra-layer interactions.

Deep Learning A-Z™ Hands-On Artificial Neural Networks(8. Part 3 - Recurrent Neural Networks)

Perma.cc archive of https://towardsdatascience.com/recurrent-neural-networks-rnn-explained-the-eli5-way-3956887e8b75 created on 2022-08-13 07:53:12.087197+00:00.

Kinect skeleton tracker is able to achieve considerable human body tracking performance in convenient and a low-cost manner. However, The tracker often captures unnatural human poses such as discontinuous and vibrated motions when self-occlusions occur. A majority of approaches tackle this problem by using multiple Kinect sensors in a workspace. Combination of the measurements from different sensors is then conducted in Kalman filter framework or optimization problem is formulated for sensor fusion. However, these methods usually require heuristics to measure reliability of measurements observed from each Kinect sensor. In this paper, we developed a method to improve Kinect skeleton using single Kinect sensor, in which supervised learning technique was employed to correct unnatural tracking motions. Specifically, deep recurrent neural networks were used for improving joint positions and velocities of Kinect skeleton, and three methods were proposed to integrate the refined positions and velocities for further enhancement. Moreover, we suggested a novel measure to evaluate naturalness of captured motions. We evaluated the proposed approach by comparison with the ground truth obtained using a commercial optical maker-based motion capture system.

Recurrent Neural Networks (RNNs) have the ability to retain memory and learn data sequences. Due to the recurrent nature of RNNs, it is sometimes hard to parallelize all its computations on conventional hardware. CPUs do not currently offer large parallelism, while GPUs offer limited parallelism due to sequential components of RNN models. In this paper we present a hardware implementation of Long-Short Term Memory (LSTM) recurrent network on the programmable logic Zynq 7020 FPGA from Xilinx. We implemented a RNN with $2$ layers and $128$ hidden units in hardware and it has been tested using a character level language model. The implementation is more than $21\times$ faster than the ARM CPU embedded on the Zynq 7020 FPGA. This work can potentially evolve to a RNN co-processor for future mobile devices.

This paper discusses models for dialogue state tracking using recurrent neural networks (RNN). We present experiments on the standard dialogue state tracking (DST) dataset, DSTC2. On the one hand, RNN models became the state of the art models in DST, on the other hand, most state-of-the-art models are only turn-based and require dataset-specific preprocessing (e.g. DSTC2-specific) in order to achieve such results. We implemented two architectures which can be used in incremental settings and require almost no preprocessing. We compare their performance to the benchmarks on DSTC2 and discuss their properties. With only trivial preprocessing, the performance of our models is close to the state-of- the-art results.

Recurrent Neural Networks (RNNs) produce state-of-art performance on many machine learning tasks but their demand on resources in terms of memory and computational power are often high. Therefore, there is a great interest in optimizing the computations performed with these models especially when considering development of specialized low-power hardware for deep networks. One way of reducing the computational needs is to limit the numerical precision of the network weights and biases. This has led to different proposed rounding methods which have been applied so far to only Convolutional Neural Networks and Fully-Connected Networks. This paper addresses the question of how to best reduce weight precision during training in the case of RNNs. We present results from the use of different stochastic and deterministic reduced precision training methods applied to three major RNN types which are then tested on several datasets. The results show that the weight binarization methods do not work with the RNNs. However, the stochastic and deterministic ternarization, and pow2-ternarization methods gave rise to low-precision RNNs that produce similar and even higher accuracy on certain datasets therefore providing a path towards training more efficient implementations of RNNs in specialized hardware.

A recurrent neural network (RNN) is a universal approximator of dynamical systems, whose performance often depends on sensitive hyperparameters. Tuning of such hyperparameters may be difficult and, typically, based on a trial-and-error approach. In this work, we adopt a graph-based framework to interpret and characterize the internal RNN dynamics. Through this insight, we are able to design a principled unsupervised method to derive configurations with maximized performances, in terms of prediction error and memory capacity. In particular, we propose to model time series of neurons activations with the recently introduced horizontal visibility graphs, whose topological properties reflect important dynamical features of the underlying dynamic system. Successively, each graph becomes a layer of a larger structure, called multiplex. We show that topological properties of such a multiplex reflect important features of RNN dynamics and are used to guide the tuning procedure. To validate the proposed method, we consider a class of RNNs called echo state networks. We perform experiments and discuss results on several benchmarks and real-world dataset of call data records.

Partially observed control problems are a challenging aspect of reinforcement learning. We extend two related, model-free algorithms for continuous control -- deterministic policy gradient and stochastic value gradient -- to solve partially observed domains using recurrent neural networks trained with backpropagation through time. We demonstrate that this approach, coupled with long-short term memory is able to solve a variety of physical control problems exhibiting an assortment of memory requirements. These include the short-term integration of information from noisy sensors and the identification of system parameters, as well as long-term memory problems that require preserving information over many time steps. We also demonstrate success on a combined exploration and memory problem in the form of a simplified version of the well-known Morris water maze task. Finally, we show that our approach can deal with high-dimensional observations by learning directly from pixels. We find that recurrent deterministic and stochastic policies are able to learn similarly good solutions to these tasks, including the water maze where the agent must learn effective search strategies.

Many natural language processing applications use language models to generate text. These models are typically trained to predict the next word in a sequence, given the previous words and some context such as an image. However, at test time the model is expected to generate the entire sequence from scratch. This discrepancy makes generation brittle, as errors may accumulate along the way. We address this issue by proposing a novel sequence level training algorithm that directly optimizes the metric used at test time, such as BLEU or ROUGE. On three different tasks, our approach outperforms several strong baselines for greedy generation. The method is also competitive when these baselines employ beam search, while being several times faster.

We investigate sequence machine learning techniques on raw radio signal time-series data. By applying deep recurrent neural networks we learn to discriminate between several application layer traffic types on top of a constant envelope modulation without using an expert demodulation algorithm. We show that complex protocol sequences can be learned and used for both classification and generation tasks using this approach.

In this paper, we systematically analyze the connecting architectures of recurrent neural networks (RNNs). Our main contribution is twofold: first, we present a rigorous graph-theoretic framework describing the connecting architectures of RNNs in general. Second, we propose three architecture complexity measures of RNNs: (a) the recurrent depth, which captures the RNN's over-time nonlinear complexity, (b) the feedforward depth, which captures the local input-output nonlinearity (similar to the "depth" in feedforward neural networks (FNNs)), and (c) the recurrent skip coefficient which captures how rapidly the information propagates over time. We rigorously prove each measure's existence and computability. Our experimental results show that RNNs might benefit from larger recurrent depth and feedforward depth. We further demonstrate that increasing recurrent skip coefficient offers performance boosts on long term dependency problems.

Recurrent neural networks (RNNs) are powerful and effective for processing sequential data. However, RNNs are usually considered "black box" models whose internal structure and learned parameters are not interpretable. In this paper, we propose an interpretable RNN based on the sequential iterative soft-thresholding algorithm (SISTA) for solving the sequential sparse recovery problem, which models a sequence of correlated observations with a sequence of sparse latent vectors. The architecture of the resulting SISTA-RNN is implicitly defined by the computational structure of SISTA, which results in a novel stacked RNN architecture. Furthermore, the weights of the SISTA-RNN are perfectly interpretable as the parameters of a principled statistical model, which in this case include a sparsifying dictionary, iterative step size, and regularization parameters. In addition, on a particular sequential compressive sensing task, the SISTA-RNN trains faster and achieves better performance than conventional state-of-the-art black box RNNs, including long-short term memory (LSTM) RNNs.

Machine learning models are frequently used to solve complex security problems, as well as to make decisions in sensitive situations like guiding autonomous vehicles or predicting financial market behaviors. Previous efforts have shown that numerous machine learning models were vulnerable to adversarial manipulations of their inputs taking the form of adversarial samples. Such inputs are crafted by adding carefully selected perturbations to legitimate inputs so as to force the machine learning model to misbehave, for instance by outputting a wrong class if the machine learning task of interest is classification. In fact, to the best of our knowledge, all previous work on adversarial samples crafting for neural network considered models used to solve classification tasks, most frequently in computer vision applications. In this paper, we contribute to the field of adversarial machine learning by investigating adversarial input sequences for recurrent neural networks processing sequential data. We show that the classes of algorithms introduced previously to craft adversarial samples misclassified by feed-forward neural networks can be adapted to recurrent neural networks. In a experiment, we show that adversaries can craft adversarial sequences misleading both categorical and sequential recurrent neural networks.

The problem where a tropical cyclone intensifies dramatically within a short period of time is known as rapid intensification. This has been one of the major challenges for tropical weather forecasting. Recurrent neural networks have been promising for time series problems which makes them appropriate for rapid intensification. In this paper, recurrent neural networks are used to predict rapid intensification cases of tropical cyclones from the South Pacific and South Indian Ocean regions. A class imbalanced problem is encountered which makes it very challenging to achieve promising performance. A simple strategy was proposed to include more positive cases for detection where the false positive rate was slightly improved. The limitations of building an efficient system remains due to the challenges of addressing the class imbalance problem encountered for rapid intensification prediction. This motivates further research in using innovative machine learning methods.

Training recurrent neural networks to model long term dependencies is difficult. Hence, we propose to use external linguistic knowledge as an explicit signal to inform the model which memories it should utilize. Specifically, external knowledge is used to augment a sequence with typed edges between arbitrarily distant elements, and the resulting graph is decomposed into directed acyclic subgraphs. We introduce a model that encodes such graphs as explicit memory in recurrent neural networks, and use it to model coreference relations in text. We apply our model to several text comprehension tasks and achieve new state-of-the-art results on all considered benchmarks, including CNN, bAbi, and LAMBADA. On the bAbi QA tasks, our model solves 15 out of the 20 tasks with only 1000 training examples per task. Analysis of the learned representations further demonstrates the ability of our model to encode fine-grained entity information across a document.

In this paper, we investigate the associative memory in recurrent neural networks, based on the model of evolving neural networks proposed by Nolfi, Miglino and Parisi. Experimentally developed network has highly asymmetric synaptic weights and dilute connections, quite different from those of the Hopfield model. Some results on the effect of learning efficiency on the evolution are also presented.

This paper presents a model based on Deep Learning algorithms of LSTM and GRU for facilitating an anomaly detection in Large Hadron Collider superconducting magnets. We used high resolution data available in Post Mortem database to train a set of models and chose the best possible set of their hyper-parameters. Using Deep Learning approach allowed to examine a vast body of data and extract the fragments which require further experts examination and are regarded as anomalies. The presented method does not require tedious manual threshold setting and operator attention at the stage of the system setup. Instead, the automatic approach is proposed, which achieves according to our experiments accuracy of 99%. This is reached for the largest dataset of 302 MB and the following architecture of the network: single layer LSTM, 128 cells, 20 epochs of training, look_back=16, look_ahead=128, grid=100 and optimizer Adam. All the experiments were run on GPU Nvidia Tesla K80

We introduce LL-RNNs (Log-Linear RNNs), an extension of Recurrent Neural Networks that replaces the softmax output layer by a log-linear output layer, of which the softmax is a special case. This conceptually simple move has two main advantages. First, it allows the learner to combat training data sparsity by allowing it to model words (or more generally, output symbols) as complex combinations of attributes without requiring that each combination is directly observed in the training data (as the softmax does). Second, it permits the inclusion of flexible prior knowledge in the form of a priori specified modular features, where the neural network component learns to dynamically control the weights of a log-linear distribution exploiting these features. We conduct experiments in the domain of language modelling of French, that exploit morphological prior knowledge and show an important decrease in perplexity relative to a baseline RNN. We provide other motivating iillustrations, and finally argue that the log-linear and the neural-network components contribute complementary strengths to the LL-RNN: the LL aspect allows the model to incorporate rich prior knowledge, while the NN aspect, according to the "representation learning" paradigm, allows the model to discover novel combination of characteristics.

A lecture notes style review of the equilibrium statistical mechanics of recurrent neural networks with discrete and continuous neurons (e.g. Ising, coupled-oscillators). To be published in the Handbook of Biological Physics (North-Holland). Accompanied by a similar review (part II) dealing with the dynamics.

For most people, watching a brief video and describing what happened (in words) is an easy task. For machines, extracting the meaning from video pixels and generating a sentence description is a very complex problem. The goal of my research is to develop models that can automatically generate natural language (NL) descriptions for events in videos. As a first step, this proposal presents deep recurrent neural network models for video to text generation. I build on recent deep machine learning approaches to develop video description models using a unified deep neural network with both convolutional and recurrent structure. This technique treats the video domain as another language and takes a machine translation approach using the deep network to translate videos to text. In my initial approach, I adapt a model that can learn on images and captions to transfer knowledge from this auxiliary task to generate descriptions for short video clips. Next, I present an end-to-end deep network that can jointly model a sequence of video frames and a sequence of words. The second part of the proposal outlines a set of models to significantly extend work in this area. Specifically, I propose techniques to integrate linguistic knowledge from plain text corpora; and attention methods to focus on objects and track their interactions to generate more diverse and accurate descriptions. To move beyond short video clips, I also outline models to process multi-activity movie videos, learning to jointly segment and describe coherent event sequences. I propose further extensions to take advantage of movie scripts and subtitle information to generate richer descriptions.

Recurrent neural networks (RNNs) have been used extensively and with increasing success to model various types of sequential data. Much of this progress has been achieved through devising recurrent units and architectures with the flexibility to capture complex statistics in the data, such as long range dependency or localized attention phenomena. However, while many sequential data (such as video, speech or language) can have highly variable information flow, most recurrent models still consume input features at a constant rate and perform a constant number of computations per time step, which can be detrimental to both speed and model capacity. In this paper, we explore a modification to existing recurrent units which allows them to learn to vary the amount of computation they perform at each step, without prior knowledge of the sequence's time structure. We show experimentally that not only do our models require fewer operations, they also lead to better performance overall on evaluation tasks.

We propose a deep learning model for identifying structure within experiment narratives in scientific literature. We take a sequence labeling approach to this problem, and label clauses within experiment narratives to identify the different parts of the experiment. Our dataset consists of paragraphs taken from open access PubMed papers labeled with rhetorical information as a result of our pilot annotation. Our model is a Recurrent Neural Network (RNN) with Long Short-Term Memory (LSTM) cells that labels clauses. The clause representations are computed by combining word representations using a novel attention mechanism that involves a separate RNN. We compare this model against LSTMs where the input layer has simple or no attention and a feature rich CRF model. Furthermore, we describe how our work could be useful for information extraction from scientific literature.

In this paper, we propose a deep neural network architecture for object recognition based on recurrent neural networks. The proposed network, called ReNet, replaces the ubiquitous convolution+pooling layer of the deep convolutional neural network with four recurrent neural networks that sweep horizontally and vertically in both directions across the image. We evaluate the proposed ReNet on three widely-used benchmark datasets; MNIST, CIFAR-10 and SVHN. The result suggests that ReNet is a viable alternative to the deep convolutional neural network, and that further investigation is needed.

Understanding the theoretical foundations of how memories are encoded and retrieved in neural populations is a central challenge in neuroscience. A popular theoretical scenario for modeling memory function is the attractor neural network scenario, whose prototype is the Hopfield model. The model has a poor storage capacity, compared with the capacity achieved with perceptron learning algorithms. Here, by transforming the perceptron learning rule, we present an online learning rule for a recurrent neural network that achieves near-maximal storage capacity without an explicit supervisory error signal, relying only upon locally accessible information. The fully-connected network consists of excitatory binary neurons with plastic recurrent connections and non-plastic inhibitory feedback stabilizing the network dynamics; the memory patterns are presented online as strong afferent currents, producing a bimodal distribution for the neuron synaptic inputs. Synapses corresponding to active inputs are modified as a function of the value of the local fields with respect to three thresholds. Above the highest threshold, and below the lowest threshold, no plasticity occurs. In between these two thresholds, potentiation/depression occurs when the local field is above/below an intermediate threshold. We simulated and analyzed a network of binary neurons implementing this rule and measured its storage capacity for different sizes of the basins of attraction. The storage capacity obtained through numerical simulations is shown to be close to the value predicted by analytical calculations. We also measured the dependence of capacity on the strength of external inputs. Finally, we quantified the statistics of the resulting synaptic connectivity matrix, and found that both the fraction of zero weight synapses and the degree of symmetry of the weight matrix increase with the number of stored patterns.

Recurrent neural networks (RNNs) were recently proposed for the session-based recommendation task. The models showed promising improvements over traditional recommendation approaches. In this work, we further study RNN-based models for session-based recommendations. We propose the application of two techniques to improve model performance, namely, data augmentation, and a method to account for shifts in the input data distribution. We also empirically study the use of generalised distillation, and a novel alternative model that directly predicts item embeddings. Experiments on the RecSys Challenge 2015 dataset demonstrate relative improvements of 12.8% and 14.8% over previously reported results on the Recall@20 and Mean Reciprocal Rank@20 metrics respectively.

The transcription of handwritten text on images is one task in machine learning and one solution to solve it is using multi-dimensional recurrent neural networks (MDRNN) with connectionist temporal classification (CTC). The RNNs can contain special units, the long short-term memory (LSTM) cells. They are able to learn long term dependencies but they get unstable when the dimension is chosen greater than one. We defined some useful and necessary properties for the one-dimensional LSTM cell and extend them in the multi-dimensional case. Thereby we introduce several new cells with better stability. We present a method to design cells using the theory of linear shift invariant systems. The new cells are compared to the LSTM cell on the IFN/ENIT and Rimes database, where we can improve the recognition rate compared to the LSTM cell. So each application where the LSTM cells in MDRNNs are used could be improved by substituting them by the new developed cells.

Bibliography: p. [14-16]

A big challenge in algorithmic composition is to devise a model that is both easily trainable and able to reproduce the long-range temporal dependencies typical of music. Here we investigate how artificial neural networks can be trained on a large corpus of melodies and turned into automated music composers able to generate new melodies coherent with the style they have been trained on. We employ gated recurrent unit networks that have been shown to be particularly efficient in learning complex sequential activations with arbitrary long time lags. Our model processes rhythm and melody in parallel while modeling the relation between these two features. Using such an approach, we were able to generate interesting complete melodies or suggest possible continuations of a melody fragment that is coherent with the characteristics of the fragment itself.

Recurrent neural networks have shown excellent performance in many applications, however they require increased complexity in hardware or software based implementations. The hardware complexity can be much lowered by minimizing the word-length of weights and signals. This work analyzes the fixed-point performance of recurrent neural networks using a retrain based quantization method. The quantization sensitivity of each layer in RNNs is studied, and the overall fixed-point optimization results minimizing the capacity of weights while not sacrificing the performance are presented. A language model and a phoneme recognition examples are used.

Recent approaches based on artificial neural networks (ANNs) have shown promising results for short-text classification. However, many short texts occur in sequences (e.g., sentences in a document or utterances in a dialog), and most existing ANN-based systems do not leverage the preceding short texts when classifying a subsequent one. In this work, we present a model based on recurrent neural networks and convolutional neural networks that incorporates the preceding short texts. Our model achieves state-of-the-art results on three different datasets for dialog act prediction.

We utilize machine learning models which are based on recurrent neural networks to optimize dynamical decoupling (DD) sequences. DD is a relatively simple technique for suppressing the errors in quantum memory for certain noise models. In numerical simulations, we show that with minimum use of prior knowledge and starting from random sequences, the models are able to improve over time and eventually output DD-sequences with performance better than that of the well known DD-families. Furthermore, our algorithm is easy to implement in experiments to find solutions tailored to the specific hardware, as it treats the figure of merit as a black box.

Leveraging large historical data in electronic health record (EHR), we developed Doctor AI, a generic predictive model that covers observed medical conditions and medication uses. Doctor AI is a temporal model using recurrent neural networks (RNN) and was developed and applied to longitudinal time stamped EHR data from 260K patients over 8 years. Encounter records (e.g. diagnosis codes, medication codes or procedure codes) were input to RNN to predict (all) the diagnosis and medication categories for a subsequent visit. Doctor AI assesses the history of patients to make multilabel predictions (one label for each diagnosis or medication category). Based on separate blind test set evaluation, Doctor AI can perform differential diagnosis with up to 79% recall@30, significantly higher than several baselines. Moreover, we demonstrate great generalizability of Doctor AI by adapting the resulting models from one institution to another without losing substantial accuracy.

Headline generation for spoken content is important since spoken content is difficult to be shown on the screen and browsed by the user. It is a special type of abstractive summarization, for which the summaries are generated word by word from scratch without using any part of the original content. Many deep learning approaches for headline generation from text document have been proposed recently, all requiring huge quantities of training data, which is difficult for spoken document summarization. In this paper, we propose an ASR error modeling approach to learn the underlying structure of ASR error patterns and incorporate this model in an Attentive Recurrent Neural Network (ARNN) architecture. In this way, the model for abstractive headline generation for spoken content can be learned from abundant text data and the ASR data for some recognizers. Experiments showed very encouraging results and verified that the proposed ASR error model works well even when the input spoken content is recognized by a recognizer very different from the one the model learned from.

In image labeling, local representations for image units are usually generated from their surrounding image patches, thus long-range contextual information is not effectively encoded. In this paper, we introduce recurrent neural networks (RNNs) to address this issue. Specifically, directed acyclic graph RNNs (DAG-RNNs) are proposed to process DAG-structured images, which enables the network to model long-range semantic dependencies among image units. Our DAG-RNNs are capable of tremendously enhancing the discriminative power of local representations, which significantly benefits the local classification. Meanwhile, we propose a novel class weighting function that attends to rare classes, which phenomenally boosts the recognition accuracy for non-frequent classes. Integrating with convolution and deconvolution layers, our DAG-RNNs achieve new state-of-the-art results on the challenging SiftFlow, CamVid and Barcelona benchmarks.

In this work we explore a straightforward variational Bayes scheme for Recurrent Neural Networks. Firstly, we show that a simple adaptation of truncated backpropagation through time can yield good quality uncertainty estimates and superior regularisation at only a small extra computational cost during training. Secondly, we demonstrate how a novel kind of posterior approximation yields further improvements to the performance of Bayesian RNNs. We incorporate local gradient information into the approximate posterior to sharpen it around the current batch statistics. This technique is not exclusive to recurrent neural networks and can be applied more widely to train Bayesian neural networks. We also empirically demonstrate how Bayesian RNNs are superior to traditional RNNs on a language modelling benchmark and an image captioning task, as well as showing how each of these methods improve our model over a variety of other schemes for training them. We also introduce a new benchmark for studying uncertainty for language models so future methods can be easily compared.

The assessment document at the time of study program accreditation shows performance achievements that will have an impact on the development of the study program in the future. The description in the assessment document contains unstructured data, making it difficult to identify target indicators. Apart from that, the number of Indonesian-based assessment documents is quite large, and there has been no research on these assessment documents. Therefore, this research aims to classify and predict target indicator categories into 4 categories: deficient, enough, good, and very. Learning testing of the Indonesian language assessment sentence classification model using recurrent neural networks-long short-term memory (RNN-LSTM) using 5 layers and 3 parameters produces performance with an accuracy value of 94.24% and a loss of 10%. In the evaluation with the Adamax optimizer, it had a high level of accuracy, namely 79%, followed by stochastic gradient descent (SGD) of 78%. For the Adam optimizer, Adadelta, and root mean squared propagation (RMSProp) have an accuracy rate of 77%.

The developed work in this paper is a part of the detection and identification of faults in systems by modern techniques of artificial intelligence. In a first step we have developed amulti-layer perceptron (MLP), type neural network to detect shunt faults and shading phenomenon in photovoltaic (PV) systems, and in the second part of the work we developed anotherrecurrent neural network (RNN) type network in order to identify single and combined faults in PV systems. The results obtained clearly show the performance of the networks developed for the rapid detection of the appearance of faults with the estimation of their times as well as the robust decision to identify the type of faults in the PV system.

A perturbative method is developed for calculating the effects of recurrent synaptic interactions between neurons embedded in a network. A series expansion is constructed that converges for networks with noisy membrane potential and weak synaptic connectivity. The terms of the series can be interpreted as loops of interactions between neurons, so the technique is called a loop-expansion. A diagrammatic method is introduced that allows for construction of analytic expressions for the parameter dependencies of the spike probability function and correlation functions. An analytic expression is obtained to predict the effect of the surrounding network on a neuron during an intracellular current injection. The analytic results are compared with simulations to test the range of their validity and significant effects of the the recurrent connections in network are accurately predicted by the loop-expansion.

Clinical medical data, especially in the intensive care unit (ICU), consist of multivariate time series of observations. For each patient visit (or episode), sensor data and lab test results are recorded in the patient's Electronic Health Record (EHR). While potentially containing a wealth of insights, the data is difficult to mine effectively, owing to varying length, irregular sampling and missing data. Recurrent Neural Networks (RNNs), particularly those using Long Short-Term Memory (LSTM) hidden units, are powerful and increasingly popular models for learning from sequence data. They effectively model varying length sequences and capture long range dependencies. We present the first study to empirically evaluate the ability of LSTMs to recognize patterns in multivariate time series of clinical measurements. Specifically, we consider multilabel classification of diagnoses, training a model to classify 128 diagnoses given 13 frequently but irregularly sampled clinical measurements. First, we establish the effectiveness of a simple LSTM network for modeling clinical data. Then we demonstrate a straightforward and effective training strategy in which we replicate targets at each sequence step. Trained only on raw time series, our models outperform several strong baselines, including a multilayer perceptron trained on hand-engineered features.

The current paper proposes a novel neural network model for recognizing visually perceived human actions. The proposed multiple spatio-temporal scales recurrent neural network (MSTRNN) model is derived by introducing multiple timescale recurrent dynamics to the conventional convolutional neural network model. One of the essential characteristics of the MSTRNN is that its architecture imposes both spatial and temporal constraints simultaneously on the neural activity which vary in multiple scales among different layers. As suggested by the principle of the upward and downward causation, it is assumed that the network can develop meaningful structures such as functional hierarchy by taking advantage of such constraints during the course of learning. To evaluate the characteristics of the model, the current study uses three types of human action video dataset consisting of different types of primitive actions and different levels of compositionality on them. The performance of the MSTRNN in testing with these dataset is compared with the ones by other representative deep learning models used in the field. The analysis of the internal representation obtained through the learning with the dataset clarifies what sorts of functional hierarchy can be developed by extracting the essential compositionality underlying the dataset.

Pixel-level labelling tasks, such as semantic segmentation, play a central role in image understanding. Recent approaches have attempted to harness the capabilities of deep learning techniques for image recognition to tackle pixel-level labelling tasks. One central issue in this methodology is the limited capacity of deep learning techniques to delineate visual objects. To solve this problem, we introduce a new form of convolutional neural network that combines the strengths of Convolutional Neural Networks (CNNs) and Conditional Random Fields (CRFs)-based probabilistic graphical modelling. To this end, we formulate mean-field approximate inference for the Conditional Random Fields with Gaussian pairwise potentials as Recurrent Neural Networks. This network, called CRF-RNN, is then plugged in as a part of a CNN to obtain a deep network that has desirable properties of both CNNs and CRFs. Importantly, our system fully integrates CRF modelling with CNNs, making it possible to train the whole deep network end-to-end with the usual back-propagation algorithm, avoiding offline post-processing methods for object delineation. We apply the proposed method to the problem of semantic image segmentation, obtaining top results on the challenging Pascal VOC 2012 segmentation benchmark.

Large-scale recurrent networks have drawn increasing attention recently because of their capabilities in modeling a large variety of real-world phenomena and physical mechanisms. This paper studies how to identify all authentic connections and estimate system parameters of a recurrent network, given a sequence of node observations. This task becomes extremely challenging in modern network applications, because the available observations are usually very noisy and limited, and the associated dynamical system is strongly nonlinear. By formulating the problem as multivariate sparse sigmoidal regression, we develop simple-to-implement network learning algorithms, with rigorous convergence guarantee in theory, for a variety of sparsity-promoting penalty forms. A quantile variant of progressive recurrent network screening is proposed for efficient computation and allows for direct cardinality control of network topology in estimation. Moreover, we investigate recurrent network stability conditions in Lyapunov's sense, and integrate such stability constraints into sparse network learning. Experiments show excellent performance of the proposed algorithms in network topology identification and forecasting.

We consider the method of Reduction of Dissipativity Domain to prove global Lyapunov stability of Discrete Time Recurrent Neural Networks. The standard and advanced criteria for Absolute Stability of these essentially nonlinear systems produce rather weak results. The method mentioned above is proved to be more powerful. It involves a multi-step procedure with maximization of special nonconvex functions over polytopes on every step. We derive conditions which guarantee an existence of at most one point of local maximum for such functions over every hyperplane. This nontrivial result is valid for wide range of neuron transfer functions.

Protein secondary structure prediction is an important problem in bioinformatics. Inspired by the recent successes of deep neural networks, in this paper, we propose an end-to-end deep network that predicts protein secondary structures from integrated local and global contextual features. Our deep architecture leverages convolutional neural networks with different kernel sizes to extract multiscale local contextual features. In addition, considering long-range dependencies existing in amino acid sequences, we set up a bidirectional neural network consisting of gated recurrent unit to capture global contextual features. Furthermore, multi-task learning is utilized to predict secondary structure labels and amino-acid solvent accessibility simultaneously. Our proposed deep network demonstrates its effectiveness by achieving state-of-the-art performance, i.e., 69.7% Q8 accuracy on the public benchmark CB513, 76.9% Q8 accuracy on CASP10 and 73.1% Q8 accuracy on CASP11. Our model and results are publicly available.

We investigate a recently proposed model for cortical computation which performs relational inference. It consists of several interconnected, structurally equivalent populations of leaky integrate-and-fire (LIF) neurons, which are trained in a self-organized fashion with spike-timing dependent plasticity (STDP). Despite its robust learning dynamics, the model is susceptible to a problem typical for recurrent networks which use a correlation based (Hebbian) learning rule: if trained with high learning rates, the recurrent connections can cause strong feedback loops in the network dynamics, which lead to the emergence of attractor states. This causes a strong reduction in the number of representable patterns and a decay in the inference ability of the network. As a solution, we introduce a conceptually very simple "wake-sleep" algorithm: during the wake phase, training is executed normally, while during the sleep phase, the network "dreams" samples from its generative model, which are induced by random input. This process allows us to activate the attractor states in the network, which can then be unlearned effectively by an anti-Hebbian mechanism. The algorithm allows us to increase learning rates up to a factor of ten while avoiding clustering, which allows the network to learn several times faster. Also for low learning rates, where clustering is not an issue, it improves convergence speed and reduces the final inference error.