arXiv research

Recent studies on automatic neural architectures search have demonstrated significant performance, competitive to or even better than hand-crafted neural architectures. However, most of the existing network architecture tend to use residual, parallel structures and concatenation block between shallow and deep features …

mGRN improves multivariate time series prediction by managing marginal and joint memories.

LSTMs and GRUs are the most common recurrent neural network architectures used to solve temporal sequence problems. The two architectures have differing data flows dealing with a common component called the cell state (also referred to as the memory). We attempt to enhance the memory by presenting a modification that w…

In neural architecture search (NAS), the space of neural network architectures is automatically explored to maximize predictive accuracy for a given task. Despite the success of recent approaches, most existing methods cannot be directly applied to large scale problems because of their prohibitive computational complex…

Many engineers wish to deploy modern neural networks in memory-limited settings; but the development of flexible methods for reducing memory use is in its infancy, and there is little knowledge of the resulting cost-benefit. We propose structural model distillation for memory reduction using a strategy that produces a …

A method to reduce memory usage in NAS by pruning the search space.

Recurrent neural networks can learn complex transduction problems that require maintaining and actively exploiting a memory of their inputs. Such models traditionally consider memory and input-output functionalities indissolubly entangled. We introduce a novel recurrent architecture based on the conceptual separation b…

Learning to remember long sequences remains a challenging task for recurrent neural networks. Register memory and attention mechanisms were both proposed to resolve the issue with either high computational cost to retain memory differentiability, or by discounting the RNN representation learning towards encoding shorte…

Improved recurrent neural networks learn long-term dependencies through multi-scale memory.

Serenity optimizes neural network execution for edge devices by scheduling with optimal memory footprint.

V-HMN integrates memory mechanisms for improved image recognition.

Study compares LSTM and ANN architectures for forex prediction, finding ANN more efficient.

We first pose the Unsupervised Progressive Learning (UPL) problem: an online representation learning problem in which the learner observes a non-stationary and unlabeled data stream, learning a growing number of features that persist over time even though the data is not stored or replayed. To solve the UPL problem we …

Memory split advantage: thinner networks outperform a single wide network.

Paper proposes DAC-ML, a cognitive architecture that learns quickly from few episodes.

New findings on how convolutional architectures approximate time series data.

A new memory-efficient Adam variant reduces second moments when feasible.

In this paper, the echo state network (ESN) memory capacity, which represents the amount of input data an ESN can store, is analyzed for a new type of deep ESNs. In particular, two deep ESN architectures are studied. First, a parallel deep ESN is proposed in which multiple reservoirs are connected in parallel allowing …

Neural architecture search (NAS) has a great impact by automatically designing effective neural network architectures. However, the prohibitive computational demand of conventional NAS algorithms (e.g. $10^4$ GPU hours) makes it difficult to \emph{directly} search the architectures on large-scale tasks (e.g. ImageNet).…

Neural networks powered with external memory simulate computer behaviors. These models, which use the memory to store data for a neural controller, can learn algorithms and other complex tasks. In this paper, we introduce a new memory to store weights for the controller, analogous to the stored-program memory in modern…

Different neural network (NN) architectures have different advantages. Convolutional neural networks (CNNs) achieved enormous success in computer vision, while recurrent neural networks (RNNs) gained popularity in speech recognition. It is not known which type of NN architecture is the best fit for classification of co…

The recent advent of `Internet of Things' (IOT) has increased the demand for enabling AI-based edge computing. This has necessitated the search for efficient implementations of neural networks in terms of both computations and storage. Although extreme quantization has proven to be a powerful tool to achieve significan…

To understand the fundamental trade-offs between training stability, temporal dynamics and architectural complexity of recurrent neural networks~(RNNs), we directly analyze RNN architectures using numerical methods of ordinary differential equations~(ODEs). We define a general family of RNNs--the ODERNNs--by relating t…

Owing to their ability to both effectively integrate information over long time horizons and scale to massive amounts of data, self-attention architectures have recently shown breakthrough success in natural language processing (NLP), achieving state-of-the-art results in domains such as language modeling and machine t…

PETRA enables parallel training of deep models with reversible architectures.

GOBO compresses 99.9% of BERT model parameters to 3 bits, improving inference efficiency.

Metagenomic studies have increasingly utilized sequencing technologies in order to analyze DNA fragments found in environmental samples.One important step in this analysis is the taxonomic classification of the DNA fragments. Conventional read classification methods require large databases and vast amounts of memory to…

Develops hyperparameter transfer methods for Dense Associative Memories.

We design and study a Contextual Memory Tree (CMT), a learning memory controller that inserts new memories into an experience store of unbounded size. It is designed to efficiently query for memories from that store, supporting logarithmic time insertion and retrieval operations. Hence CMT can be integrated into existi…

Momentum ResNets improve ResNets' memory efficiency.

Due to its perceptual limitations, an agent may have too little information about the state of the environment to act optimally. In such cases, it is important to keep track of the observation history to uncover hidden state. Recent deep reinforcement learning methods use recurrent neural networks (RNN) to memorize pas…

We introduce a new deep learning architecture for predicting price movements from limit order books. This architecture uses a causal convolutional network for feature extraction in combination with masked self-attention to update features based on relevant contextual information. This architecture is shown to significa…

In many learning situations, resources at inference time are significantly more constrained than resources at training time. This paper studies a general paradigm, called Differentiable ARchitecture Compression (DARC), that combines model compression and architecture search to learn models that are resource-efficient a…

Hydra boosts efficiency for long-context reasoning in resource-constrained settings.

The explosion in workload complexity and the recent slow-down in Moore's law scaling call for new approaches towards efficient computing. Researchers are now beginning to use recent advances in machine learning in software optimizations, augmenting or replacing traditional heuristics and data structures. However, the s…

In distributed reinforcement learning, it is common to exchange the experience memory of each agent and thereby collectively train their local models. The experience memory, however, contains all the preceding state observations and their corresponding policies of the host agent, which may violate the privacy of the ag…

This work tackles exploding inverses in INNs, revealing and mitigating their numerical non-invertibility.

Modelling long-term dependencies is a challenge for recurrent neural networks. This is primarily due to the fact that gradients vanish during training, as the sequence length increases. Gradients can be attenuated by transition operators and are attenuated or dropped by activation functions. Canonical architectures lik…

Neural network tackles continual learning with neuromodulation and local error signals.

To be effective in sequential data processing, Recurrent Neural Networks (RNNs) are required to keep track of past events by creating memories. While the relation between memories and the network's hidden state dynamics was established over the last decade, previous works in this direction were of a predominantly descr…

New methods improve neural network extrapolation to long sequences.

Meta-learning agents excel at rapidly learning new tasks from open-ended task distributions; yet, they forget what they learn about each task as soon as the next begins. When tasks reoccur - as they do in natural environments - metalearning agents must explore again instead of immediately exploiting previously discover…

This paper presents a novel framework for predicting shot location and type in tennis. Inspired by recent neuroscience discoveries we incorporate neural memory modules to model the episodic and semantic memory components of a tennis player. We propose a Semi Supervised Generative Adversarial Network architecture that c…

Memory is an important aspect of intelligence and plays a role in many deep reinforcement learning models. However, little progress has been made in understanding when specific memory systems help more than others and how well they generalize. The field also has yet to see a prevalent consistent and rigorous approach f…

The extension of deep learning towards temporal data processing is gaining an increasing research interest. In this paper we investigate the properties of state dynamics developed in successive levels of deep recurrent neural networks (RNNs) in terms of short-term memory abilities. Our results reveal interesting insigh…

Recent dialogue approaches operate by reading each word in a conversation history, and aggregating accrued dialogue information into a single state. This fixed-size vector is not expandable and must maintain a consistent format over time. Other recent approaches exploit an attention mechanism to extract useful informat…

Deep feedforward and recurrent networks have achieved impressive results in many perception and language processing applications. This success is partially attributed to architectural innovations such as convolutional and long short-term memory networks. The main motivation for these architectural innovations is that t…

We consider a novel question answering (QA) task where the machine needs to read from large streaming data (long documents or videos) without knowing when the questions will be given, which is difficult to solve with existing QA methods due to their lack of scalability. To tackle this problem, we propose a novel end-to…