EXPLORE / 06
Terms index
Specialized AI and software-development terms with concise explanations
Entity records
4377 records · 20/110
- 761
SimReady Asset
3D models pre-configured with physical properties (e.g., collision, mass) that are ready for immediate use in simulation environ
Website pending - 762
Secure Runtime
An isolated environment for agents to execute code or call tools, featuring permission controls, sandboxing, and audit logs to p
Website pending - 763
Agent Grounding
The process of anchoring an agent's reasoning and actions to specific contexts or authoritative data sources (e.g., enterprise k
Website pending - 764
CUDA-X
A collection of high-performance computing libraries built on CUDA by NVIDIA, covering AI, data science, graphics, and physics s
Website pending - 765
Sparse Direct Solver
An algorithm that solves matrix equations with mostly zero elements via exact factorization rather than iterative approximation,
Website pending - 766
Distributed AI Network
An infrastructure architecture leveraging dispersed edge nodes (e.g., telecom base stations) to collaboratively provide AI compu
Website pending - 767
Edge AI
Technology that runs AI algorithms locally on devices rather than in the cloud, emphasizing low latency, privacy, and offline ca
Website pending - 768
Event Streaming
A data transmission pattern that continuously records discrete events for consumers to read sequentially, commonly used for buil
Website pending - 769
AI Compute Extensions (ACE)
A next-generation x86 instruction set architecture designed to accelerate core AI operations like matrix multiplication at the h
Website pending - 770
Hybrid Bonding
An advanced packaging technology achieving bumpless interconnects via direct copper-to-copper and dielectric-to-dielectric bondi
Website pending - 771
De Novo Design
The process of constructing new molecules with specific structures and functions from scratch, based on physicochemical principl
Website pending - 772
Protein Binder
Engineered small proteins designed to specifically recognize and bind to target molecules (such as toxins or receptors), often u
Website pending - 773
Ontology Knowledge Graph
A data structure based on formal ontologies defining entity relationships, supporting semantic reasoning, cross-domain reuse, an
Website pending - 774
Model Context Protocol Server
A standardized protocol interface allowing LLMs to securely connect to external data sources and tools for context awareness and
Website pending - 775
Hardware-Software Co-design
Tailoring underlying hardware instruction sets or accelerators to specific AI model characteristics to maximize inference effici
Website pending - 776
End-to-End Autonomous Driving
A deep learning architecture that maps sensor inputs directly to control outputs without traditional modular rule engines, enabl
Website pending - 777
CUDA Moat
The ecosystem barrier established by NVIDIA via its CUDA software platform, making it difficult for developers to migrate to oth
Website pending - 778
Code Reachability Analysis
Static or dynamic determination of which code segments may actually execute under specific paths, used to identify dead code or
Website pending - 779
Attack Surface Reduction
A security strategy that minimizes potential entry points for attackers by removing, disabling, or restricting non-essential fea
Website pending - 780
Verified Software
Software accompanied by mathematical proofs checked by computer-assisted tools (like Lean) to rule out specific classes of bugs.
Website pending - 781
Formal Semantics
A mathematical definition of the meaning of programming language constructs, serving as the foundation for proving program corre
Website pending - 782
Natural Language Instruction
A paradigm using LLMs to interpret human text intent and translate it into specific operational parameters or generation constra
Website pending - 783
Parallel Memory Pathway
A computational branch running parallel to the original attention mechanism in Transformer modules, dedicated to processing inte
Website pending - 784
Memory-Native Model
An architecture that internalizes memory writing, updating, and reading as learnable capabilities, distinct from methods relying
Website pending - 785
Agentic Evaluation
Benchmark systems assessing LLM capabilities in multi-step planning, tool invocation, and autonomous execution of complex tasks.
Website pending - 786
i-line Lithography
Lithography using the 365nm mercury i-line as the light source, commonly used in mature nodes and pattern transfer in advanced p
Website pending - 787
RDL
A metal interconnection layer used to redistribute signals between the chip and external pins, common in advanced packaging.
Website pending - 788
Cost per GPU-Hour
A key metric for evaluating the economic efficiency of LLM inference or training, representing the rental or depreciation cost r
Website pending - 789
Quantized Inference
A technique that reduces memory footprint and accelerates computation by lowering the numerical precision of model weights or ac
Website pending - 790
OLAP
Online Analytical Processing systems optimized for complex queries and multidimensional analysis over large datasets, commonly u
Website pending - 791
Text-to-SQL
Technology converting natural language queries into SQL, where accuracy heavily depends on domain-specific semantic layer constr
Website pending - 792
Agent-Native Interface
A human-computer interaction paradigm centered on multi-agent task flows that weakens or removes traditional manual editing cont
Website pending - 793
WAF
A security mechanism designed to filter and monitor HTTP/HTTPS traffic to block common web attacks such as SQL injection and XSS
Website pending - 794
Black-box Testing
A testing method that validates system functionality and security by observing input-output behavior without knowledge of intern
Website pending - 795
Thinking Mode
A decoding strategy that improves accuracy on complex problems by generating intermediate reasoning steps, often accompanied by
Website pending - 796
Source Recall
The proportion or quantity of valid information sources successfully found and returned relevant to a query in a retrieval syste
Website pending - 797
Hybrid Indexing
A technique combining keyword exact matching (inverted index) and semantic similarity search (vector index) to balance precision
Website pending - 798
MCU
A compact integrated circuit designed to govern a specific operation in an embedded system, containing processor, memory, and I/
Website pending - 799
Generative Completion
The process of synthesizing missing textures, structures, or pixel content based on the model's understanding of image semantics
Website pending - 800
Real-time Multimodal Inference
The capability to process images, audio, and sensor data simultaneously under low-latency constraints, serving as a critical tec
Website pending