Results for "neural networks"

Achieving task performance by providing a small number of examples inside the prompt without weight updates.

Chooses which experts process each token.

Reusing knowledge from a source task/domain to improve learning on a target task/domain, typically via pretrained models.

A continuous vector encoding of an item (word, image, user) such that semantic similarity corresponds to geometric closeness.

Predicts masked tokens in a sequence, enabling bidirectional context; often used for embeddings rather than generation.

A high-capacity language model trained on massive corpora, exhibiting broad generalization and emergent behaviors.

PEFT method injecting trainable low-rank matrices into layers, enabling efficient fine-tuning.

Diffusion model trained to remove noise step by step.

Model that compresses input into latent space and reconstructs it.

Assigning category labels to images.

Pixel-level separation of individual object instances.

Deep learning system for protein structure prediction.

Protecting data during network transfer and while stored; essential for ML pipelines handling sensitive data.

Architecture based on self-attention and feedforward layers; foundation of modern LLMs and many multimodal models.

Bayesian parameter estimation using the mode of the posterior distribution.

Artificially created data used to train/test models; helpful for privacy and coverage, risky if unrealistic.

The range of functions a model can represent.

Expected return of taking action in a state.

Structured graph encoding facts as entity–relation–entity triples.

Extension of convolution to graph domains using adjacency structure.

Simplified Boltzmann Machine with bipartite structure.

Graphical model expressing factorization of a probability distribution.

Models that learn to generate samples resembling training data.

Generative model that learns to reverse a gradual noise process.

Generator produces limited variety of outputs.

Changing speaker characteristics while preserving content.

Identifying speakers in audio.

Decomposing goals into sub-tasks.

Agent reasoning about future outcomes.

Storing results to reduce compute.