AI terminology
| Term . | Definition . |
|---|---|
| Algorithm bias | Systematic errors in AI models that result in unfair treatment of certain patient populations due to imbalanced training data |
| AI | The simulation of human intelligence in machines that can perform tasks such as learning, reasoning, and problem-solving |
| Data drift | Changes in data patterns over time that can reduce AI model performance, requiring continuous monitoring and updates |
| DL | A specialized ML approach using artificial neural networks with multiple layers to analyze complex data |
| Digital twin | A virtual representation of a patient using AI to simulate disease progression and treatment responses |
| Generative adversarial networks | A class of ML frameworks in which 2 neural networks, a generator and a discriminator, compete in a game-theoretic setup. The generator creates data samples (eg, images, text) that resemble real data, whereas the discriminator evaluates their authenticity. Over time, the generator improves at producing realistic outputs. |
| Generative AI | AI systems that create new content, such as text, images, or synthetic data, using models such as generative adversarial networks and transformer-based language models |
| GPT-3 | Generative pretrained transformer 3 is a multimodal large language model trained and created by OpenAI and the third in its series of GPT foundation models |
| ML | A subset of AI that enables computers to learn patterns from data and make predictions without explicit programming |
| MLDevOps | The integration of ML development and operational processes to ensure the reliable deployment of AI models in health care |
| Natural language processing | A field of AI that enables computers to understand, interpret, and generate human language |
| Neural networks | AI structures modeled after the human brain, composed of interconnected nodes (neurons) for data processing |
| Reinforcement learning | An AI approach in which an agent learns by interacting with an environment and receiving rewards or penalties |
| Supervised learning | An ML technique in which algorithms are trained on labeled datasets, mapping inputs to known outputs |
| Transformer-based large language models | A type of deep learning model based on the transformer architecture, which uses self-attention mechanisms to process and generate human-like text |
| Unsupervised learning | An ML technique that identifies patterns and structures in unlabeled data without predefined outcomes |
| Term . | Definition . |
|---|---|
| Algorithm bias | Systematic errors in AI models that result in unfair treatment of certain patient populations due to imbalanced training data |
| AI | The simulation of human intelligence in machines that can perform tasks such as learning, reasoning, and problem-solving |
| Data drift | Changes in data patterns over time that can reduce AI model performance, requiring continuous monitoring and updates |
| DL | A specialized ML approach using artificial neural networks with multiple layers to analyze complex data |
| Digital twin | A virtual representation of a patient using AI to simulate disease progression and treatment responses |
| Generative adversarial networks | A class of ML frameworks in which 2 neural networks, a generator and a discriminator, compete in a game-theoretic setup. The generator creates data samples (eg, images, text) that resemble real data, whereas the discriminator evaluates their authenticity. Over time, the generator improves at producing realistic outputs. |
| Generative AI | AI systems that create new content, such as text, images, or synthetic data, using models such as generative adversarial networks and transformer-based language models |
| GPT-3 | Generative pretrained transformer 3 is a multimodal large language model trained and created by OpenAI and the third in its series of GPT foundation models |
| ML | A subset of AI that enables computers to learn patterns from data and make predictions without explicit programming |
| MLDevOps | The integration of ML development and operational processes to ensure the reliable deployment of AI models in health care |
| Natural language processing | A field of AI that enables computers to understand, interpret, and generate human language |
| Neural networks | AI structures modeled after the human brain, composed of interconnected nodes (neurons) for data processing |
| Reinforcement learning | An AI approach in which an agent learns by interacting with an environment and receiving rewards or penalties |
| Supervised learning | An ML technique in which algorithms are trained on labeled datasets, mapping inputs to known outputs |
| Transformer-based large language models | A type of deep learning model based on the transformer architecture, which uses self-attention mechanisms to process and generate human-like text |
| Unsupervised learning | An ML technique that identifies patterns and structures in unlabeled data without predefined outcomes |
GPT-3, generative pretrained transformer 3.