{"url":"/method/memory-network","slug":"memory-network","name":"Memory Network","full_name":"Memory Network","full_name_withheld":false,"description_markdown":"A **Memory Network** provides a memory component that can be read from and written to with the inference capabilities of a neural network model. The motivation is that many neural networks lack a long-term memory component, and their existing memory component encoded by states and weights is too small and not compartmentalized enough to accurately remember facts from the past (RNNs for example, have difficult memorizing and doing tasks like copying). \r\n\r\nA memory network consists of a memory $\\textbf{m}$ (an array of objects indexed by $\\textbf{m}\\_{i}$ and four potentially learned components:\r\n\r\n- Input feature map $I$ - feature representation of the data input.\r\n- Generalization $G$ - updates old memories given the new input.\r\n- Output feature map $O$ - produces new feature map given $I$ and $G$.\r\n- Response $R$ - converts output into the desired response. \r\n\r\nGiven an input $x$ (e.g., an input character, word or sentence depending on the granularity chosen, an image or an audio signal) the flow of the model is as follows:\r\n\r\n1. Convert $x$ to an internal feature representation $I\\left(x\\right)$.\r\n2. Update memories $m\\_{i}$ given the new input: $m\\_{i} = G\\left(m\\_{i}, I\\left(x\\right), m\\right)$, $\\forall{i}$.\r\n3. Compute output features $o$ given the new input and the memory: $o = O\\left(I\\left(x\\right), m\\right)$.\r\n4. Finally, decode output features $o$ to give the final response: $r = R\\left(o\\right)$.\r\n\r\nThis process is applied at both train and test time, if there is a distinction between such phases, that\r\nis, memories are also stored at test time, but the model parameters of $I$, $G$, $O$ and $R$ are not updated. Memory networks cover a wide class of possible implementations. The components $I$, $G$, $O$ and $R$ can potentially use any existing ideas from the machine learning literature.\r\n\r\nImage Source: [Adrian Colyer](https://blog.acolyer.org/2016/03/10/memory-networks/)","description_state":"present","introduced_year":null,"introduced_by":{"title":null,"paper":null,"first_author":null,"n_authors":0,"url_abs":null,"archive_paper_url":null},"source":{"url":"http://arxiv.org/abs/1410.3916v11","title":"Memory Networks","url_on_a_paper_host":true},"code_snippet_url":"https://github.com/aykutaaykut/Memory-Networks","code_snippet_url_on_a_code_host":true,"categories":[{"area":"General","area_id":"general","collection":"Working Memory Models","url":"/methods/category/working-memory-models","pwc_aliases":[]}],"n_papers_tagged":445,"archive_num_papers":null,"papers_newest_first":[{"paper":null,"title":"Deep learning for predicting hauling fleet production capacity under uncertainties in open pit mines using real and simulated data","date":"2025-06-04","arxiv_id":"2506.04296","n_code_links":0,"syntology":null},{"paper":null,"title":"A novel Neural-ODE model for the state of health estimation of lithium-ion battery using charging curve","date":"2025-05-09","arxiv_id":"2505.05803","n_code_links":0,"syntology":null},{"paper":null,"title":"Using Machine Learning and Neural Networks to Analyze and Predict Chaos in Multi-Pendulum and Chaotic Systems","date":"2025-04-18","arxiv_id":"2504.13453","n_code_links":0,"syntology":null},{"paper":null,"title":"Predicting Driver's Perceived Risk: a Model Based on Semi-Supervised Learning Strategy","date":"2025-04-17","arxiv_id":"2504.12665","n_code_links":0,"syntology":null},{"paper":null,"title":"Oscillatory Associative Memory with Exponential Capacity","date":"2025-04-04","arxiv_id":"2504.03102","n_code_links":0,"syntology":null},{"paper":null,"title":"Deep Representation Learning for Unsupervised Clustering of Myocardial Fiber Trajectories in Cardiac Diffusion Tensor Imaging","date":"2025-04-02","arxiv_id":"2504.01953","n_code_links":0,"syntology":null},{"paper":null,"title":"Incremental capacity-based multi-feature fusion model for predicting state-of-health of lithium-ion batteries","date":"2025-03-31","arxiv_id":"2503.23858","n_code_links":0,"syntology":null},{"paper":null,"title":"Emotion Detection in Twitter Messages Using Combination of Long Short-Term Memory and Convolutional Deep Neural Networks","date":"2025-03-26","arxiv_id":"2503.20163","n_code_links":0,"syntology":null},{"paper":"/paper/savid-spectravista-aesthetic-vision","title":"SaViD: Spectravista Aesthetic Vision Integration for Robust and Discerning 3D Object Detection in Challenging Environments","date":"2025-03-26","arxiv_id":"2503.20614","n_code_links":1,"syntology":null},{"paper":null,"title":"Coupling deep and handcrafted features to assess smile genuineness","date":"2025-03-20","arxiv_id":"2503.16128","n_code_links":0,"syntology":null},{"paper":"/paper/one-shot-medical-video-object-segmentation","title":"One-Shot Medical Video Object Segmentation via Temporal Contrastive Memory Networks","date":"2025-03-19","arxiv_id":"2503.14979","n_code_links":1,"syntology":null},{"paper":null,"title":"Long-VMNet: Accelerating Long-Form Video Understanding via Fixed Memory","date":"2025-03-17","arxiv_id":"2503.13707","n_code_links":0,"syntology":null},{"paper":"/paper/unctrack-reliable-visual-object-tracking-with","title":"UncTrack: Reliable Visual Object Tracking with Uncertainty-Aware Prototype Memory Network","date":"2025-03-17","arxiv_id":"2503.12888","n_code_links":1,"syntology":null},{"paper":null,"title":"Statistical Study of Sensor Data and Investigation of ML-based Calibration Algorithms for Inexpensive Sensor Modules: Experiments from Cape Point","date":"2025-03-09","arxiv_id":"2503.13487","n_code_links":0,"syntology":null},{"paper":null,"title":"Prediction of Halo Coronal Mass Ejections Using SDO/HMI Vector Magnetic Data Products and a Transformer Model","date":"2025-03-05","arxiv_id":"2503.03237","n_code_links":0,"syntology":null},{"paper":null,"title":"A Hybrid CNN-Transformer Model for Heart Disease Prediction Using Life History Data","date":"2025-03-03","arxiv_id":"2503.02124","n_code_links":0,"syntology":null},{"paper":null,"title":"R$^3$Mem: Bridging Memory Retention and Retrieval via Reversible Compression","date":"2025-02-21","arxiv_id":"2502.15957","n_code_links":0,"syntology":null},{"paper":"/paper/a-mem-agentic-memory-for-llm-agents","title":"A-MEM: Agentic Memory for LLM Agents","date":"2025-02-17","arxiv_id":"2502.12110","n_code_links":4,"syntology":{"ran":6,"of":10,"unverified":4,"pointer_only":0}},{"paper":null,"title":"Aerial Reliable Collaborative Communications for Terrestrial Mobile Users via Evolutionary Multi-Objective Deep Reinforcement Learning","date":"2025-02-09","arxiv_id":"2502.05824","n_code_links":0,"syntology":null},{"paper":null,"title":"Large Memory Network for Recommendation","date":"2025-02-08","arxiv_id":"2502.05558","n_code_links":0,"syntology":null},{"paper":null,"title":"A Deep Learning Framework Integrating CNN and BiLSTM for Financial Systemic Risk Analysis and Prediction","date":"2025-02-07","arxiv_id":"2502.06847","n_code_links":0,"syntology":null},{"paper":null,"title":"Subtle variations in stiff dimensions of brain networks account for individual differences in cognitive ability","date":"2025-01-31","arxiv_id":"2501.19106","n_code_links":0,"syntology":null},{"paper":null,"title":"ISAM-MTL: Cross-subject multi-task learning model with identifiable spikes and associative memory networks","date":"2025-01-30","arxiv_id":"2501.18089","n_code_links":0,"syntology":null},{"paper":null,"title":"A two-stage dual-task learning strategy for early prediction of pathological complete response to neoadjuvant chemotherapy for breast cancer using dynamic contrast-enhanced magnetic resonance images","date":"2025-01-28","arxiv_id":"2502.00051","n_code_links":0,"syntology":null},{"paper":null,"title":"Leveraging Video Vision Transformer for Alzheimer's Disease Diagnosis from 3D Brain MRI","date":"2025-01-27","arxiv_id":"2501.15733","n_code_links":0,"syntology":null},{"paper":null,"title":"Gaze Prediction as a Function of Eye Movement Type and Individual Differences","date":"2024-12-31","arxiv_id":"2501.00597","n_code_links":0,"syntology":null},{"paper":null,"title":"Protein Structure Prediction in the 3D HP Model Using Deep Reinforcement Learning","date":"2024-12-29","arxiv_id":"2412.20329","n_code_links":0,"syntology":null},{"paper":null,"title":"SCKF-LSTM Based Trajectory Tracking for Electricity-Gas Integrated Energy System","date":"2024-12-24","arxiv_id":"2412.18357","n_code_links":0,"syntology":null},{"paper":null,"title":"Distributed solar generation forecasting using attention-based deep neural networks for cloud movement prediction","date":"2024-11-17","arxiv_id":"2411.10921","n_code_links":0,"syntology":null},{"paper":"/paper/livos-light-video-object-segmentation-with","title":"LiVOS: Light Video Object Segmentation with Gated Linear Matching","date":"2024-11-05","arxiv_id":"2411.02818","n_code_links":1,"syntology":null}],"papers_shown":30,"tasks":[{"task":"/task/question-answering","name":"Question Answering","papers":37},{"task":"/task/time-series-1","name":"Time Series","papers":28},{"task":"/task/classification","name":"General Classification","papers":26},{"task":"/task/semantic-segmentation","name":"Semantic Segmentation","papers":26},{"task":"/task/prediction","name":"Prediction","papers":23},{"task":"/task/time-series","name":"Time Series Analysis","papers":23},{"task":"/task/decoder","name":"Decoder","papers":21},{"task":"/task/video-semantic-segmentation","name":"Video Semantic Segmentation","papers":21},{"task":"/task/object","name":"Object","papers":20},{"task":"/task/sentence","name":"Sentence","papers":20},{"task":"/task/video-object-segmentation","name":"Video Object Segmentation","papers":19},{"task":"/task/classification-1","name":"Classification","papers":18},{"task":"/task/segmentation","name":"Segmentation","papers":18},{"task":"/task/language-modelling","name":"Language Modelling","papers":17},{"task":"/task/retrieval","name":"Retrieval","papers":16},{"task":"/task/sentiment-analysis","name":"Sentiment Analysis","papers":15},{"task":"/task/language-modeling","name":"Language Modeling","papers":14},{"task":"/task/deep-learning","name":"Deep Learning","papers":13},{"task":"/task/decision-making","name":"Decision Making","papers":12},{"task":"/task/representation-learning","name":"Representation Learning","papers":11}],"tasks_shown":20,"n_tasks":386,"usage_by_year":[{"year":"2014","papers":1},{"year":"2015","papers":2},{"year":"2016","papers":11},{"year":"2017","papers":29},{"year":"2018","papers":47},{"year":"2019","papers":74},{"year":"2020","papers":52},{"year":"2021","papers":79},{"year":"2022","papers":45},{"year":"2023","papers":42},{"year":"2024","papers":37},{"year":"2025","papers":25}],"row_source":"embedded","archive":{"source":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","archive_url":"https://paperswithcode.com/method/memory-network"},"syntology_read_at":"2026-09-24T18:15:14+00:00"}