Seonglae ChoFeature learning, Feature Embedding
Almost things are related to Dimension Reduction
Representation Learning notion
Vector Database
For production, when using pgVector in an RDBMS rather than a vector database, simply adding an embedding column to an existing table is a bad practice. This is because if you finetune or change the model, you have to regenerate embeddings in all tables, which is very unstable and difficult to manage during migration. Therefore, it's better practice to have a separate embedding table that references the original id table like a junction table. While you can have separate tables per entity, sharing the same table doesn't pose a Polymorphic FK problem in this case. When the embedding version changes, you can simply populate new rows and migrations can flow continuously. Additionally, you can delete stale entries after migration, making it much more stable. This solution also doesn't require column expansion even when multiple embedding types are added per table, and embeddings from multiple tables can be shared like ReBAC, which is an advantage.
Embeddings are underrated
Machine learning (ML) has the potential to advance the state of the
art in technical writing. No, I’m not talking about text generation models
like Claude, Gemini, LLaMa, GPT, etc. The ML technology that might end up
having the biggest impact on technical writing is embeddings.
https://technicalwriting.dev/data/embeddings.html
Feature learning
In machine learning, feature learning or representation learning is a set of techniques that allows a system to automatically discover the representations needed for feature detection or classification from raw data. This replaces manual feature engineering and allows a machine to both learn the features and use them to perform a specific task.
https://en.wikipedia.org/wiki/Feature_learning
Papers with Code - Representation Learning
**Representation Learning** is a process in machine learning where algorithms extract meaningful patterns from raw data to create representations that are easier to understand and process. These representations can be designed for interpretability, reveal hidden features, or be used for transfer learning. They are valuable across many fundamental machine learning tasks like [image classification](/task/image-classification) and [retrieval](/task/image-retrieval). Deep neural networks can be considered representation learning models that typically encode information which is projected into a different subspace. These representations are then usually passed on to a linear classifier to, for instance, train a classifier. Representation learning can be divided into: - **Supervised representation learning**: learning representations on task A using annotated data and used to solve task B - **Unsupervised representation learning**: learning representations on a task in an unsupervised way (label-free data). These are then used to address downstream tasks and reducing the need for annotated data when learning news tasks. Powerful models like [GPT](/method/gpt) and [BERT](/method/bert) leverage unsupervised representation learning to tackle language tasks. More recently, [self-supervised learning (SSL)](/task/self-supervised-learning) is one of the main drivers behind unsupervised representation learning in fields like computer vision and NLP. Here are some additional readings to go deeper on the task: - [Representation Learning: A Review and New Perspectives](/paper/representation-learning-a-review-and-new) - Bengio et al. (2012) - [A Few Words on Representation Learning](https://sthalles.github.io/a-few-words-on-representation-learning/) - Thalles Silva <span style="color:grey; opacity: 0.6">( Image credit: [Visualizing and Understanding Convolutional Networks](https://arxiv.org/pdf/1311.2901.pdf) )</span>
https://paperswithcode.com/task/representation-learning
