15,762 technical terms and definitions
neural architecture, search nas automl, darts enas gpu nas passed rl, neural architecture search nas & automl, neural architecture search, architecture search nas, architecture search, search nas
Neural Architecture Search: NAS & AutoML Introduction & Motivation Neural Architecture Search: automatically design architectures. NAS: search space, strategy, evaluation.…
one shot nas, weight sharing nas, supernet architecture search, efficient nas darts
Neural Architecture Search (NAS) Efficiency Methods is a set of techniques that reduce the computational cost of automated architecture discovery from thousands of GPU-days…
neural architecture search (nas) and automated design space exploration, neural architecture search, architecture search (nas), search (nas) automated, (nas) automated design
Neural Architecture Search (NAS) and Automated Design Space Exploration Introduction & Motivation Neural Architecture Search (NAS) and Automated Design Space Exploration addresses…
Neural Architecture Search with Reinforcement Gradients Introduction & Motivation Neural Architecture Search with Reinforcement Gradients addresses a central problem in automated…
Neural Architecture Search with Reinforcement Gradients (Extended Study) Introduction & Motivation Neural Architecture Search with Reinforcement Gradients (Extended Study)…
Neural Architecture Search with Reinforcement Gradients (Extended Study) Introduction & Motivation Neural Architecture Search with Reinforcement Gradients (Extended Study)…
neural architecture
Neural Architecture Transfer is a NAS technique that transfers architecture knowledge across different tasks or datasets — reusing architectures or search strategies discovered on…
Neural Artistic Generation Introduction & Motivation Neural Artistic Generation addresses a central problem in generative generation and artistic-quality-metrics synthesis: how…
multimodal ai
Neural Articulation is modeling articulated object or body motion using learnable kinematic-aware neural representations - It supports controllable animation and pose-consistent…
audio & speech
Neural Beamforming is beamforming pipelines where neural networks estimate masks, covariance, or beam weights - It integrates data-driven learning with spatial filtering…
model optimization
Neural Cache is a memory-augmented mechanism that reuses recent activations or context to improve inference efficiency - It can reduce repeated computation and improve local…
recommendation systems
Neural CF is a neural collaborative-filtering framework that replaces linear interaction functions with deep nonlinear modeling - User and item embeddings are combined through…
intel neural chat, neural chat model
Neural Chat is a 7B parameter language model developed by Intel as a fine-tune of Mistral-7B, aligned using Direct Preference Optimization (DPO) and optimized to showcase…
reinforcement learning
Neural Circuit Policies (NCPs) are sparse, interpretable recurrent neural network architectures — derived from Liquid Time-Constant (LTC) networks and wired to resemble biological…
ncp, reinforcement learning
Neural Circuit Policies (NCPs) are compact, interpretable control architectures using liquid time constant neurons organized as wiring-constrained circuits, achieving robust…
multimodal ai
Neural Codec is a learned compression framework that encodes signals into compact discrete or continuous latent representations - It supports efficient multimodal storage…
Neural Collaborative Filtering Introduction & Motivation Neural Collaborative Filtering addresses a central problem in recsys systems and (target) optimization: how to Build…
Neural Collaborative Filtering with GMF and MLP Introduction & Motivation Neural Collaborative Filtering with GMF and MLP addresses a central problem in recsys systems…
structured prediction
Neural constituency parsing is constituency parsing methods that score spans or trees with neural representations - Neural encoders provide contextual token embeddings used…
neural architecture
Neural CDEs are a neural network architecture that parameterizes the response function of a controlled differential equation with a neural network — dz_t = f_theta(z_t) , dX_t…
nlp
Neural data-to-text is the approach of using neural network models for generating natural language from structured data — employing deep learning architectures (Transformers…
neural architecture search
Neural Encoding is learned embedding of architecture graphs produced by neural encoders for NAS tasks. - It aims to capture structural similarity more effectively than…
edge ai
Neural Engine is Apple's dedicated hardware accelerator for on-device machine learning, integrated into A-series (iPhone/iPad) and M-series (Mac/iPad Pro) chips — providing…
neural architecture search
Neural fabrics is a neural-architecture framework that embeds many scale and depth pathways in a unified fabric graph - Information flows through interconnected processing paths…
time series models
Neural Hawkes process is a neural temporal point-process model that learns event intensity dynamics from historical event sequences - Recurrent latent states summarize history…
3d vision
Neural implicit functions is the coordinate-based neural models that represent signals or geometry as continuous functions rather than discrete grids - they provide flexible…
computer vision
Neural implicit surfaces are a way of representing 3D surfaces using neural networks — learning continuous surface representations as implicit functions (SDF, occupancy) encoded…
neural interface electrode array yield prediction, neural interface electrode, interface electrode array, electrode array yield, array yield prediction, neural interface
Neural Interface Electrode Array Yield Prediction with Machine Learning Introduction & Motivation Neural Interface Electrode Array Yield Prediction with Machine Learning addresses…
sequence to sequence translation, transformer translation model, attention alignment translation, multilingual translation model
Neural machine translation (NMT) uses an end-to-end trained neural network, rather than the separately engineered components (translation model, language model, reordering model)…
multimodal ai
Neural Mesh is a mesh representation whose geometry or texture parameters are optimized with neural methods - It combines explicit topology control with learnable high-quality…
3d vision
Neural mesh representation is the hybrid 3D modeling approach that combines mesh topology with neural features for geometry and appearance - it merges explicit surface control…
reasoning
Neural Module Composition is the architectural paradigm where neural network layouts are dynamically assembled at inference time by selecting and connecting specialized…
what is a neural network, neural networks, artificial neural network, ann, neural net, how do neural networks work
A neural network is a machine-learning model that maps inputs to outputs by passing data through layers of simple units, each computing a weighted sum of its inputs followed by…
tpu, npu, systolic array, ai chip, hardware ai inference, tensor processing unit
Neural Network Accelerators are the specialized hardware processors designed to perform the matrix multiply-accumulate (MAC) operations that dominate neural network inference…
neural network architectures, mlp, cnn, rnn, transformer, gnn, architecture design
Neural network architecture is the structural design of a learned computational graph: its layers, connections, width, depth, state, activation, and information flow. Architecture…
ml driven rtl generation, ai circuit generation, automated hdl synthesis, learning based logic synthesis
Neural Network Synthesis is the emerging paradigm of using deep learning models to directly generate hardware descriptions, optimize logic circuits, and synthesize chip designs…
ml compiler, graph optimization, tvm compiler
Neural Network Compilers are the software systems that transform high-level model definitions (PyTorch/TensorFlow graphs) into optimized low-level code for specific hardware…
neural network compiler optimization and hardware mapping, neural network compiler, network compiler optimization, compiler optimization hardware, optimization hardware mapping
Neural Network Compiler Optimization and Hardware Mapping Introduction & Motivation Neural Network Compiler Optimization and Hardware Mapping addresses a central problem…
online distillation, co distillation, mutual learning, collaborative training
Online Distillation and Co-Distillation is the training paradigm where multiple neural networks teach each other simultaneously during training — unlike traditional knowledge…
control theory
Neural Network Dynamics Models are data-driven models that use neural networks to learn the dynamics of physical or manufacturing systems — replacing first-principles equations…
nngp, theory
NNGP (Neural Network Gaussian Process) is a theoretical result showing that infinitely wide neural networks with random weights converge to Gaussian Processes — the distribution…
weight initialization, xavier glorot, kaiming he, training convergence
Neural Network Initialization Strategies — Setting the Foundation for Successful Training Weight initialization is a critical yet often underappreciated aspect of neural network…
neural network model compression and accuracy prediction, neural network model, network model compression, model compression accuracy, compression accuracy prediction
Neural Network Model Compression and Accuracy Prediction with Machine Learning Introduction & Motivation Neural Network Model Compression and Accuracy Prediction with Machine…
adam optimizer, learning rate schedule, gradient descent variant, optimizer training
Neural Network Optimizers are the algorithms that update model parameters to minimize the loss function during training — where the choice of optimizer (SGD, Adam, AdamW, LAMB)…
optimizer momentum weight decay, adamw optimizer training, lars lamb optimizer, optimizer convergence properties
Neural Network Optimizers are the algorithms that update model parameters based on computed gradients to minimize the training loss function — with the choice of optimizer (SGD…
chemistry ai
Neural Network Potentials (NNPs) are the preeminent architectural framework used to construct Machine Learning Force Fields, defining the total potential energy of a massive…
weight pruning, structured pruning, model sparsity
Neural network pruning trades model size and inference compute cost against accuracy, removing weights or structural units specifically identified as contributing least to…
unstructured structured pruning, magnitude pruning, lottery ticket hypothesis, sparsity neural network
Neural network pruning methodology has evolved from simple one-shot magnitude-based weight removal toward more sophisticated approaches, iterative pruning schedules, structured…
structured unstructured pruning, lottery ticket hypothesis, magnitude pruning, model compression sparsity
Neural network pruning applied to large language models faces distinct challenges from pruning smaller, task-specific networks, since LLMs' enormous parameter count and…
model sparsity, weight pruning, structured pruning, sparse neural networks
Neural network pruning combined with other compression techniques, quantization (reducing numerical precision) and knowledge distillation (training a smaller model to mimic…