Paper presents a new method to train deep neural networks with reduced memory access.
problem High computational and storage complexity of deep neural networks.
method Boolean logic minimization to remove memory access and reduce resource usage.
result Significantly lower latency and two orders of magnitude fewer computing resources.
We perform the first study of the tradeoff space of access methods and replication to support statistical analytics using first-order methods executed in the main memory of a Non-Uniform Memory Access (NUMA) machine. Statistical analytics systems differ from conventional SQL-analytics in the amount and types of memory …
One of the core tasks in multi-view learning is to capture relations among views. For sequential data, the relations not only span across views, but also extend throughout the view length to form long-term intra-view and inter-view interactions. In this paper, we present a new memory augmented neural network model that…
Developing efficient and scalable algorithms for Latent Dirichlet Allocation (LDA) is of wide interest for many applications. Previous work has developed an O(1) Metropolis-Hastings sampling method for each token. However, the performance is far from being optimal due to random accesses to the parameter matrices and fr…
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…
Study reduces memory needs for active learning with enriched queries.
problem Expensive labeling costs in active learning.
method Introduces bounded memory active learning through enriched queries, introduces lossless sample compression.
result Can learn classifiers with bounded memory and query optimality.
Proposes two techniques for efficient tail averaging in streams.
problem Efficiently averaging the last examples in a stream with low memory and constant access.
method Two techniques with low constant memory cost and access to average at every time step. Improved accuracy at higher memory cost.
result Proposes techniques for efficient tail averaging with constant memory and access at every time step.
New technique trains DNNs with fewer weights, saving memory and energy.
problem Training deep neural networks requires many weights, increasing memory and energy costs.
method Constrain weight updates to those with highest gradients, regenerating others.
result Pruned networks maintain accuracy while significantly reducing weight count and memory usage.
We present a memory augmented neural network for natural language understanding: Neural Semantic Encoders. NSE is equipped with a novel memory update rule and has a variable sized encoding memory that evolves over time and maintains the understanding of input sequences through read}, compose and write operations. NSE c…
Quantum memory limits set by relativity theory.
problem Quantum memory efficiency and relativity constraints.
method Relativistic quantum field theory and Lieb-Robinson bounds.
result Quantum memory capacity is limited by fundamental physics.
New method optimizes memory usage in neural networks, improving sequential learning.
problem Current memory models in neural networks waste memory and computation.
method Formulated an optimization problem to maximize information storage, introduced Cached Uniform Writing.
result Proved Cached Uniform Writing optimizes memory usage and outperforms other methods.
Many machine learning approaches are characterized by information constraints on how they interact with the training data. These include memory and sequential access constraints (e.g. fast first-order methods to solve stochastic optimization problems); communication constraints (e.g. distributed learning); partial acce…
Study shows Transformer and Neural GPU are Turing complete without external memory.
problem Exploring computational power of modern neural network architectures.
method Analyzing computational properties of Transformer and Neural GPU.
result Transformer and Neural GPU are Turing complete without external memory.
SliceOut speeds up deep learning training without sacrificing accuracy.
problem Frequent model re-training and large model training workloads in deep learning.
method SliceOut uses dropout-inspired scheme to drop contiguous sets of units at random, leveraging GPU memory layout.
result 10-40% speedups and memory reduction with minimal accuracy loss.
New algorithm SAGA++ improves on SAGA for faster convergence in stochastic batch size methods.
problem Optimizing convergence rate in stochastic variance reduction methods with batch size.
method Proposes SAGA++ algorithm with optimal average batch size considering cache/disk IO effects.
result SAGA++ outperforms SAGA and other solvers on real datasets.
Big T-Rex solves FDR-controlled sparse regression on laptops with millions of variables.
problem Scalable FDR-controlled variable selection for high-dimensional data.
method Early terminated random experiments with memory-mapping and permutation-based dummy generation.
result Solves FDR-controlled Lasso problems with 5 million variables on a laptop in 30 minutes.
CICLAD efficiently mines frequent closed itemsets from data streams with minimal memory usage.
problem Mining frequent closed itemsets from data streams is resource-intensive.
method CICLAD is an intersection-based sliding-window FCI miner that optimizes memory usage while maintaining performance.
result CICLAD achieves significantly lower memory footprint compared to existing methods.
Study shows BNN inference accelerators are vulnerable to soft errors, causing significant misclassification.
problem Vulnerability of BNN inference accelerators to soft errors affecting image classification accuracy.
method Fault injection experiments on CIFAR-10 and MNIST datasets to assess impact of soft errors.
result Accuracy of image classifier can drop by up to 76.70% in worst-case scenarios.
This paper enhances ML algorithms by improving data locality and reducing redundancy.
problem Improving performance of machine learning algorithms with complex data.
method Exploiting data locality and reuse in memory hierarchies of modern processors.
result Efficient implementation of machine learning algorithms can be achieved by reusing computation results.
Paper uses machine learning to optimize UAV deployment for traffic offloading.
problem Optimizing UAV deployment for efficient traffic offloading from ground BSs.
method LSTM for traffic prediction, KEG algorithm for service area determination, multi-access techniques comparison.
result RSMA reduces up to 24% total power consumption compared to conventional methods.
Study shows how to balance memory and learning efficiency in continual learning.
problem Balancing memory and learning efficiency in continual learning.
method Structural regularization with Hessian-based regularization.
result Structural regularization improves statistical performance at the cost of increased memory complexity.
A method to improve sequential learning by keeping past data errors in check.
problem Challenges in sequential learning with Gaussian processes due to accumulating errors.
method Memory-based dual sparse variational Gaussian processes.
result Improves accuracy in inference and learning for various applications.
Proposes a plastic neural memory model for better anomaly detection.
problem Static attention mechanisms limit NMNs in anomaly detection.
method Introduces dynamic connection weights for improved knowledge retrieval.
result Outperforms state-of-the-art in three medical anomaly detection tasks.
Quadratic memory is essential for optimal convex optimization queries.
problem Optimal query complexity for convex optimization and feasibility problems.
method Lower bounds on query complexity for convex optimization and feasibility problems.
result Center-of-mass algorithms are Pareto-optimal for both convex optimization and feasibility problems.
A new method learns from multi-modal sequences with external memory.
problem Learning new modes in a dynamic environment without prior knowledge.
method Maintains a neural episodic memory with a Dirichlet Process prior to store mode descriptors and transfers knowledge through retrieval.
result Performs continual learning favorably compared to mainstream approaches.
SAGE improves memory efficiency by selectively adding, merging, or ignoring new facts.
problem Efficiently managing new facts in agentic LLMs to avoid costly write-time reasoning.
method SAGE uses a von Mises-Fisher-based density estimator to score and route candidate facts.
result SAGE achieves the best average token-F1 on LoCoMo and reduces add-phase API cost by 3.4x on GPT-4o-mini.
Compact DNNs increase memory footprint and reduce energy efficiency.
problem Designing compact deep neural networks (DNNs) for improved energy efficiency.
method Evaluation of recently proposed compact DNNs on a Tesla P100 GPU.
result Higher number of activations and memory footprint lead to reduced energy efficiency.
Improved memory efficiency for deep neural networks.
problem Increasing model size makes deep learning less accessible.
method Inductive-bias-driven network design and manifold-oriented regularization.
result Improved performance and generalization error compared to traditional methods.
PoET-BiN reduces power consumption in neural networks on embedded devices.
problem Power inefficiency in neural network implementations on embedded platforms.
method Look-Up Table based implementation with a modified Decision Tree approach.
result Near state-of-the-art results with up to 6 orders of magnitude energy reduction.
BiQGEMM efficiently multiplies quantized DNN weights using lookup tables.
problem Efficiently multiplying quantized DNN weights on CPUs/GPUs with limited memory.
method BiQGEMM pre-computes and stores redundant intermediate results in lookup tables.
result BiQGEMM achieves lower overall computations and higher performance.
FibQuant improves KV-cache compression for long-context inference.
problem Memory traffic bottleneck in long-context inference due to KV cache growth.
method Introduces FibQuant, a universal vector quantizer that combines Beta-quantile radii, Fibonacci/Roberts-Kronecker directions, and Lloyd-Max refinement.
result FibQuant achieves high compression rates with minimal loss in attention cosine similarity.
Animals execute goal-directed behaviours despite the limited range and scope of their sensors. To cope, they explore environments and store memories maintaining estimates of important information that is not presently available. Recently, progress has been made with artificial intelligence (AI) agents that learn to per…
Gradient descent is efficient for solving feasibility problems with minimal memory and queries.
problem Finding a point in a given set using a memory-constrained algorithm with a separation oracle.
method Oracle complexity lower bounds for gradient descent and other algorithms.
result Gradient descent is Pareto-optimal in the oracle complexity/memory tradeoff for feasibility problems.
Accelerator boosts energy efficiency for MANNs on FPGAs.
problem Efficiently running MANNs on accelerators designed for other NNs.
method Data flow architecture, inference thresholding.
result Higher energy efficiency compared to NVIDIA GPU.
Improves energy efficiency of neuromorphic hardware by optimizing memory organization and encoding schemes.
problem Energy inefficiency in neuromorphic hardware, especially in digital accelerators.
method Synthesized controller and memory for different encoding schemes, introduced functional encoding for structured connectivity.
result Functional encoding offers a 58% reduction in energy for weight updates in convolutional layers.
New methods improve neural network extrapolation to long sequences.
problem Neural networks struggle with extrapolation to very long or adversarial sequences.
method Activation binning and localized differentiable memory architecture.
result No extrapolation errors detected within memory constraints.
Parrot learns optimal cache replacement policies using imitation learning.
problem Improving cache hit rates in complex access patterns.
method Imitation learning approach using Belady's oracle policy.
result Parrot increases cache hit rates by 61% on a web search benchmark.
Machine reading using differentiable reasoning models has recently shown remarkable progress. In this context, End-to-End trainable Memory Networks, MemN2N, have demonstrated promising performance on simple natural language based reasoning tasks such as factual reasoning and basic deduction. However, other tasks, namel…
Performing signal processing tasks on compressive measurements of data has received great attention in recent years. In this paper, we extend previous work on compressive dictionary learning by showing that more general random projections may be used, including sparse ones. More precisely, we examine compressive K-mean…
New fair PCA method using streaming algorithms with statistical guarantees.
problem Perform PCA while ensuring fair projected distributions.
method Formulated new notion PAFO learnability, proposed fair noisy power method (FNPM) for memory efficiency.
result First statistical guarantee for fair PCA in streaming setting.
SmartExchange trades memory for computation in neural networks.
problem Energy-efficient inference of deep neural networks.
method Algorithm-hardware co-design framework integrating sparsification, decomposition, and quantization.
result SmartExchange reduces energy consumption and improves performance.
Trojans can be inserted into deep neural networks without altering training data.
problem Inserting malicious code into deep learning models without changing training data.
method Patches model parameters in memory to achieve desired behavior on specific inputs.
result Efficient patches can induce trojan behavior with minimal risk of detection.
Convolutional sparse representations are a form of sparse representation with a structured, translation invariant dictionary. Most convolutional dictionary learning algorithms to date operate in batch mode, requiring simultaneous access to all training images during the learning process, which results in very high memo…
The brain optimizes memory by forgetting what's predictable, improving generalization.
problem Memory consolidation struggles with representational drift, semanticisation, and offline replay.
method Proposes predictive forgetting as a mechanism to optimize generalization by reducing complexity.
result Predictive forgetting improves information-theoretic generalization bounds on stored representations.
Muon outperforms GD in associative memory learning by balancing frequency components.
problem Training dynamics and scaling behavior of Muon in associative memory learning.
method Study of Muon in a linear associative memory model with softmax retrieval and hierarchical frequency spectrum over query-answer pairs.
result Muon achieves exponential speedup over GD in noiseless case and superior scaling efficiency in noisy case.
Study language generation with limited memory, showing different impacts on achievable densities and convergence.
problem Language generation with bounded memory constraints.
method Analyzed memoryless generators, sliding windows, and adaptive past examples; revisited identification in the limit.
result Achievable densities and convergence properties differ based on the size of the target language collection.
AIDEL improves scalability of learned indexes in storage systems.
problem Expensive retraining and heavy inter-model dependency in learned indexes limit scalability.
method Construct different linear regression models based on data distribution, making them independent and easier to partition.
result AIDEL improves insertion performance by about 2x and comparable lookup performance.
The study proves necessary conditions for robust decision-making in uncertain environments.
problem Conditions for robust decision-making in uncertain environments.
method Quantitative selection theorems and binary betting decisions.
result World models, belief-like memory, and persistent variables are necessary for strong task performance.