AI Glossary: Letter "K"

K-Means Clustering is an unsupervised machine learning algorithm that partitions a dataset into K distinct, non-overlapping clusters by assigning each data point to its nearest centroid.

K-Nearest Neighbors (KNN) is a simple, non-parametric supervised learning algorithm used for classification and regression, which predicts target labels by looking at the majority vote of K closest neighboring coordinates.

Kahneman-Tversky Optimization (KTO) is an alignment objective based on behavioral economics (Prospect Theory) that updates policy weights using unpaired binary feedback (desirable/undesirable labels) rather than paired preferences, optimizing model utility relative to a status quo baseline.

A Knowledge Graph is a structured database representing a network of real-world entities (nodes) and their semantic relations (edges), allowing systems to query logical context.

Knowledge Graph Integration is the process of combining structured, relational facts from a Knowledge Graph with the generative capabilities of an LLM. This provides models with access to explicit, verifiable relationship mappings, enhancing facts grounding.

A KV Cache (Key-Value Cache) is an inference-time optimization storing the computed Key and Value attention tensors of past tokens to prevent redundant recalculations in autoregressive decoding.

KV Cache Eviction is a memory management technique that removes less important key-value states from GPU memory during long text generation. This prevents out-of-memory errors and keeps sequence processing fast.