Generative memory helps AI learn new tasks without forgetting old ones.
problem Catastrophic forgetting in lifelong reinforcement learning.
method Generative memory model for task-agnostic latent space.
result Generative memory can improve lifelong reinforcement learning performance.
New memory allocation scheme improves image generation performance.
problem Improving episodic and semantic memory representation in neural networks.
method Developed a hierarchical latent variable model with differentiable, locally block allocated latent memory.
result Improved conditional likelihood values on various datasets.
Generative diffusion models mimic biological memory networks, encoding associative dynamics in deep neural weights.
problem Understanding long-term memory mechanisms in neuroscience and AI.
method Interpreting generative diffusion models as energy-based models and comparing them to Hopfield networks.
result Generative diffusion models can encode associative dynamics of Hopfield networks in deep neural weights.
LEMN improves memory networks for learning from streaming data.
problem Efficiently processing large data streams in memory-augmented neural networks.
method LEMN uses a RNN-based retention agent to learn and replace less important memory entries based on their importance and historical context.
result LEMN achieves significant improvements over existing methods in learning from streaming data.
We present an end-to-end trained memory system that quickly adapts to new data and generates samples like them. Inspired by Kanerva's sparse distributed memory, it has a robust distributed reading and writing mechanism. The memory is analytically tractable, which enables optimal on-line compression via a Bayesian updat…
DMs emerge from DenseAMs, transitioning from memorization to generalization.
problem Hindered memory retrieval in DenseAMs due to spurious states.
method Examined diffusion models through the lens of DenseAMs, focusing on their generative process.
result Identified a critical phase in DMs transitioning from memorization to generalization.
MEMGAN uses memory to improve anomaly detection by isolating abnormal data.
problem Weak guarantees for detecting anomalous data in classical algorithms.
method Memory-augmented Generative Adversarial Networks (MEMGAN) with a memory module.
result MEMGAN provides strong guarantees for anomaly detection with improved reconstruction.
We discuss memory models which are based on tensor decompositions using latent representations of entities and events. We show how episodic memory and semantic memory can be realized and discuss how new memory traces can be generated from sensory input: Existing memories are the basis for perception and new memories ar…
Generative memory model avoids vanishing gradients to robustly retrieve patterns.
problem Robust retrieval of stored patterns in the presence of interference and noise.
method Training a generative distributed memory without explicitly simulating attractor dynamics, using a likelihood-based Lyapunov function.
result The model converges to correct patterns upon iterative retrieval and achieves competitive performance as a memory model and a generative model.
New memory models improve LSSVM's generalization and reduce time cost.
problem Improving LSSVM's generalization performance and reducing training time.
method Introducing MIMM and WIMM models that modify LSSVM's memory influence mechanism.
result MIMM and WIMM models outperform LSSVM in generalization and time efficiency.
The paper aims to develop a methodology to test memory systems in reinforcement learning agents.
problem Understanding and evaluating memory systems in reinforcement learning agents and their generalization.
method Construct diverse memory tasks, develop an agent architecture combining multiple memory systems, and perform ablations.
result The agent architecture can apply learned knowledge to holdout data that differs from the training data.
CloGAN uses memory replay and GAN to prevent forgetting in continual learning.
problem Catastrophic forgetting in sequential learning tasks.
method Cumulative closed-loop memory replay GAN with external regularization.
result Model performance asymptotically approaches full dataset training.
A technique identifies memoryless algorithms approximating memory-dependent optimization methods.
problem Understanding how memory in optimization algorithms affects loss and generalization.
method Introducing a general technique to replace past iterates with the current one and adding a correction term.
result Lion does not have the same implicit anti-regularization as AdamW, explaining its better generalization performance.
A new memory system handles non-stationary environments by self-sizing and retaining memories.
problem Non-stationary environments where memories arrive over time and must be distinguished from noise.
method A self-sizing continual associative memory that generalizes Hopfield's network, handling adaptation and novelty.
result The memory system grows to the intrinsic memory demand of the environment and retains memories without forgetting.
Low-precision quantization improves kernel approximation under memory constraints.
problem Training kernel approximation methods efficiently with limited memory.
method Low-precision quantization of random Fourier features (LP-RFFs).
result LP-RFFs can match the performance of full-precision RFFs and Nyström method with significantly less memory.
mGRN improves multivariate time series prediction by managing marginal and joint memories.
problem Extracting dependencies in multivariate sequential data with strong serial and cross-sectional dependencies.
method Developed a novel recurrent network architecture, Memory-Gated Recurrent Networks (mGRN), with gates for marginal and joint memories.
result mGRN consistently outperforms state-of-the-art architectures on various public datasets.
A general market model with memory is considered in terms of stochastic functional differential equations. We aim at representation formulae for the sensitivity analysis of the dependence of option prices on the memory. This implies a generalization of the concept of delta.
Memory-enhanced model predicts tennis shots based on player history.
problem Predicting shot location and type in tennis.
method Semi-supervised Generative Adversarial Network with neural memory modules.
result The model learns player-specific behavioral patterns from match data.
It is generally accepted that many time series of practical interest exhibit strong dependence, i.e., long memory. For such series, the sample autocorrelations decay slowly and log-log periodogram plots indicate a straight-line relationship. This necessitates a class of models for describing such behavior. A popular cl…
New memory effect discovered in gravitational wave behavior.
problem Understanding gravitational wave behavior in spacetimes with angular momentum.
method Mathematical analysis of Minkowski spacetime and Kerr black holes.
result Angular momentum memory effect observed at future null infinity.
APL learns from surprising observations to quickly generalize from few examples.
problem Quickly generalize from limited data for intelligent systems.
method Approximates probability distributions by remembering surprising observations in an external memory module.
result APL performs as well as state-of-the-art baselines on few-shot classification benchmarks.
Tiny episodic memory significantly improves continual learning performance.
problem Transfer knowledge between tasks in continual learning.
method Store and replay a small number of examples from previous tasks.
result A simple baseline outperforms CL approaches with episodic memory.
Tree-SMU enables strong compositional generalization in neural networks.
problem Zero-shot generalization to novel compositions of concepts.
method Tree Stack Memory Units (Tree-SMU) with Stack Memory Units (SMU).
result Tree-SMU achieves strong empirical results on mathematical reasoning benchmarks.
Accurate volatility modelling is paramount for optimal risk management practices. One stylized feature of financial volatility that impacts the modelling process is long memory explored in this paper for alternative risk measures, observed absolute and squared returns for high frequency intraday UK futures. Volatility …
Typical neural networks with external memory do not effectively separate capacity for episodic and working memory as is required for reasoning in humans. Applying knowledge gained from psychological studies, we designed a new model called Differentiable Working Memory (DWM) in order to specifically emulate human workin…
Generalizes memory and forecasting capacities for nonlinear recurrent networks with dependent inputs.
problem Understanding memory and forecasting capabilities in networks with dependent inputs.
method Formulated bounds for memory and forecasting capacities in terms of network size and input properties.
result Proved that memory capacity for linear recurrent networks with independent inputs is given by the rank of the controllability matrix.
Active-memory mechanisms can replace self-attention in Transformers, but optimal results often require both.
problem Replacing self-attention with active-memory mechanisms in Transformers.
method Evaluation of various active-memory mechanisms in a Transformer model.
result Active-memory mechanisms can achieve comparable results to self-attention for language modeling, but optimal results are often achieved by combining both mechanisms.
We study the long-term memory in diverse stock market indices and foreign exchange rates using the Detrended Fluctuation Analysis(DFA). For all daily and high-frequency market data studied, no significant long-term memory property is detected in the return series, while a strong long-term memory property is found in th…
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.
We describe a new class of learning models called memory networks. Memory networks reason with inference components combined with a long-term memory component; they learn how to use these jointly. The long-term memory can be read and written to, with the goal of using it for prediction. We investigate these models in t…
Proposes a novel SAM operator for separate item and relational memories.
problem Limited memory interactions in neural networks.
method Introduces a Self-attentive Associative Memory (SAM) operator to separate item and relational memories.
result Achieves competitive results in various tasks, including geometry, graph, reinforcement learning, and question answering.
SMT improves robotic long-horizon tasks by embedding and utilizing past observations.
problem Long-horizon tasks in partially observable environments require effective long-term memory.
method Scene Memory Transformer (SMT) embeds and uses attention to exploit spatio-temporal dependencies.
result SMT outperforms existing policies in visual navigation tasks.
Dual memory improves reinforcement learning efficiency.
problem Training inefficiency in reinforcement learning.
method Introduces a dual memory structure with a main and cache memory.
result Dual memory structure leads to higher scores in reinforcement learning environments.
NEST optimizes deep learning training by placing devices efficiently across networks and memory.
problem Inefficient device placement in distributed deep learning leads to high communication and memory overhead.
method NEST uses network-, compute-, and memory-aware dynamic programming to optimize device placement.
result NEST achieves up to 2.43 times higher throughput and better memory efficiency.
Enhanced LSTM model learns complex temporal dependencies.
problem Modeling long-term dependencies in sequential data.
method Gamma-LSTM with hierarchical memory units and gates.
result Gamma-LSTM outperforms regular and stacked LSTMs in sequence prediction.
Paper proposes efficient optimizers for large language models with fast convergence and low memory usage.
problem Designing efficient optimizers for large language models with low-memory requirements and fast convergence.
method Structured Fisher information matrix approximation and low-rank extension framework.
result New optimizers (RACS and Alice) achieve better convergence and lower memory usage than existing methods.
A novel memory network improves interpretability and training for EHR analysis.
problem Limited interpretability and training difficulty in traditional memory networks for EHR analysis.
method Proposes a hybrid memory architecture with explicit and blurred memory blocks for improved interpretability and training efficiency.
result The model performs comparably to state-of-the-art methods and enables ready interpretation of results.
Service-induced congestion in memory-constrained LLM serving
problem Service-induced congestion in memory-constrained large language model (LLM) serving
method Developing a discrete-time dynamical model of memory-constrained LLM inference
result The system converges to a unique worst-case limit cycle that is asymptotically stable outside a Lebesgue-measure-zero exact-capture set, with throughput losses as large as 50%.
Eye tracking measures ADHD-related working memory deficits.
problem Diagnosing ADHD in adults requires reliable measures of working memory capacity.
method Eye tracking technology and machine learning applied to a working memory task.
result Machine learning generated features unique to ADHD.
Meta-learning approach for fast and compressive energy-based memory models.
problem Learning associative memory with fast and compressive models for complex data.
method Meta-learning approach to energy-based memory models (EBMM) using arbitrary neural architectures.
result Demonstrated associative retrieval outperforming existing systems in reconstruction error and compression rate.
New method combines long-memory reservoirs for accurate dengue forecasting from short data.
problem Accurate dengue forecasting from short, noisy, non-stationary, and nonlinear data.
method Fractional ESN and Wavelet ESN frameworks integrating long-term memory.
result fESN and wESN outperform baselines in multiple dengue datasets and forecasting horizons.
Extended LSTMs improve volatility prediction by 20%.
problem Predicting asset price volatility with long memory.
method Extended LSTMs with multiple flexible timescales.
result Extended LSTMs outperform rough volatility predictions by 20%.
A new method reduces memory usage in neural networks by 36-81%.
problem Minimizing memory usage in neural networks during backpropagation.
method Formalized as a graph theory problem, uses dynamic programming to minimize computational overhead.
result Reduces peak memory consumption by 36-81% on various networks.
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.
ORIGAMI accelerates ML algorithms by splitting compute tasks between in-memory and off-chip accelerators.
problem Memory bandwidth bottleneck in ML processing.
method Heterogeneous in-memory accelerators and off-chip compute platform, pattern-matching for compute patterns, computation-splitting compiler.
result ORIGAMI outperforms state-of-the-art accelerators in performance and energy-efficiency.
We consider streaming, one-pass principal component analysis (PCA), in the high-dimensional regime, with limited memory. Here, p-dimensional samples are presented sequentially, and the goal is to produce the k-dimensional subspace that best approximates these points. Standard algorithms require O(p2) memory; mea…
Model learns continuously from text without forgetting.
problem Catastrophic forgetting in lifelong language learning.
method Episodic memory with sparse experience replay and local adaptation.
result Model can continuously learn from new datasets.
Paper tackles online optimization with memory and competitive control.
problem Minimizing hitting and switching costs in online optimization problems.
method Optimistic Regularized Online Balanced Descent algorithm.
result Achieves a constant, dimension-free competitive ratio.