The goal of an ML project is to build a model that (helps to) solve a business problem.
#Typical Purposes
- Automate, e.g., take action on the user’s behalf, start/stop a specific activity, etc.
- Alert or prompt, e.g., ask the user if an action should be taken; alert a sysadmin about potentially suspicious network traffic.
- Organize, e.g., sort pictures or documents in the order of similarity to a query or according to the user’s preferences.
- Annotate, e.g., add contextual annotations to displayed information, or highlight phrases relevant to the user’s task in a text.
- Extract, e.g., detect smaller pieces of relevant information in a larger input, such as named entities in the text: proper names, companies, or locations.
- Recommend, e.g., detect and show to a user highly relevant items in a large collection based on item’s content or user’s reaction to the past recommendations.
- Classify, e.g., dispatch input examples into one, or several, of a predefined set of distinctly-named groups.
- Quantify, e.g., assign a number, such as a price, to an object, such as a house.
- Synthesize, e.g., generate new text, image, sound similar to the objects in a collection.
- Answer a question, e.g., “Does this text describe that image?”, “are these images similar?”
- Transform its input, e.g., reduce its dimensionality for visualization purposes, paraphrase a long text as a short abstract, translate a sentence into another language, or augment an image by applying a filter to it.
- Detect a novelty or an anomaly, e.g., fraud detection in financial transactions or predictive maintenance in manufacturing.
Deep Learning Strategy