Stochastic Gradient Descent (SGD) variants are different methods built on the basic SGD algorithm, which is used to train machine learning models by updating their parameters step by step. These variants aim to improve performance by making the updates faster, more stable, or more accurate. Some common variants include Momentum, Adam, RMSprop, and Adagrad, each…
Category: Artificial Intelligence
Model Interpretability
Model interpretability refers to how easily a human can understand the decisions or predictions made by a machine learning model. It is about making the inner workings of a model transparent, so people can see why it made a certain choice. This is important for trust, accountability, and identifying mistakes or biases in automated systems.
Knowledge Injection
Knowledge injection is the process of adding specific information or facts into an artificial intelligence system, such as a chatbot or language model, to improve its accuracy or performance. This can be done by directly feeding the system extra data, rules, or context that it would not otherwise have known. Knowledge injection helps AI systems…
Synthetic Feature Generation
Synthetic feature generation is the process of creating new data features from existing ones to help improve the performance of machine learning models. These new features are not collected directly but are derived by combining, transforming, or otherwise manipulating the original data. This helps models find patterns that may not be obvious in the raw…
Label Noise Robustness
Label noise robustness refers to the ability of a machine learning model to perform well even when some of its training data labels are incorrect or misleading. In real-world datasets, mistakes can occur when humans or automated systems assign the wrong category or value to an example. Robust models can tolerate these errors and still…
Multi-Task Learning
Multi-task learning is a machine learning approach where a single model is trained to perform several related tasks at the same time. By learning from multiple tasks, the model can share useful information between them, often leading to better overall performance. This technique can help the model generalise better and make more efficient use of…
Neural Symbolic Integration
Neural Symbolic Integration is an approach in artificial intelligence that combines neural networks, which learn from data, with symbolic reasoning systems, which follow logical rules. This integration aims to create systems that can both recognise patterns and reason about them, making decisions based on both learned experience and clear, structured logic. The goal is to…
Robust Optimization
Robust optimisation is a method in decision-making and mathematical modelling that aims to find solutions that perform well even when there is uncertainty or variability in the input data. Instead of assuming that all information is precise, it prepares for worst-case scenarios by building in a margin of safety. This approach helps ensure that the…
Fairness-Aware Machine Learning
Fairness-Aware Machine Learning refers to developing and using machine learning models that aim to make decisions without favouring or discriminating against individuals or groups based on sensitive characteristics such as gender, race, or age. It involves identifying and reducing biases that can exist in data or algorithms to ensure fair outcomes for everyone affected by…
Out-of-Distribution Detection
Out-of-Distribution Detection is a technique used to identify when a machine learning model encounters data that is significantly different from the data it was trained on. This helps to prevent the model from making unreliable or incorrect predictions on unfamiliar inputs. Detecting these cases is important for maintaining the safety and reliability of AI systems…