os&e examples

OS&E Examples: Real-World Applications & Use Cases

By Rachel Smith
AD

Open-Set and Closed-Set problems, often abbreviated as OSE, are fundamental concepts in machine learning and statistics. They are fundamental when it comes to understanding the scope of a prediction task. In this article, we'll delve into these concepts, explore their key differences, and provide real-world examples to illustrate each.

At their core, Open-Set and Closed-Set problems differ in the scope of answers they can handle. A Closed-Set problem assumes a finite and known set of possible answers, while an Open-Set problem can handle an infinite or unknown set of potential answers.

Open-Set Problems

Open-Set problems are more dynamic and flexible, accommodating an infinite set of possible outcomes. This could include anything from predicting an unrestricted numerical value to recognizing any possible image.

AD
The EOS Model
The EOS Model
MOs, EOs, & AOs
MOs, EOs, & AOs
Ella | MH Advocate & Educator | An updated guide to OSDD! Is there anything else you’d like to add to this guide? What’s one thing you like about living with OSDD? :)... | Instagram Osdd System, Osdd-1b System, Roles In A Did System, Understanding Did Symptoms, D.i.d Symptoms, Osdd-1b System Art, Osdd Systems, Understanding Dissociation In Literature, D.i.d System
Ella | MH Advocate & Educator | An updated guide to OSDD! Is there anything else you’d like to add to this guide? What’s one thing you like about living with OSDD? :)... | Instagram Osdd System, Osdd-1b System, Roles In A Did System, Understanding Did Symptoms, D.i.d Symptoms, Osdd-1b System Art, Osdd Systems, Understanding Dissociation In Literature, D.i.d System

Consider an autonomous vehicle's speed prediction. Here, the speed of the vehicle can be any real number (e.g., 32.17 mph), making this an Open-Set problem. The algorithm must predict a continuum of possible values, not just a set of predetermined speeds.

Open-Set Problem Examples

Here are two more examples of Open-Set problems:

OSCE Exam Review
OSCE Exam Review
What's in OneSkin's FACE moisturizer?
What's in OneSkin's FACE moisturizer?
this system has osd - id
this system has osd - id
  • Speech-to-Text Conversion: The words to be recognized are not known a priori and can be any of the vast number of words in one or more languages. The algorithms need to handle an open vocabulary and can even learn new words over time.
  • Stock Market Price Prediction: Stock prices are real numbers and can take any value within a certain range. Neural networks trained on historical data can predict future prices within this infinite range.

Challenges of Open-Set Problems

Despite their flexibility, Open-Set problems come with unique challenges. One significant challenge is that the disconnect between the training data and the real-world data can lead to unseen, unpredictable tests during deployment.

For instance, in the stock market example, a sudden global event could dramatically change the range of possible stock prices, leading to a misprediction. Handling and parameterizing the 'unknown unknowns' is a significant challenge in Open-Set problems.

AD

Closed-Set Problems

In contrast, Closed-Set problems have a finite set of predetermined answers. These sets can be discrete or ordinal, like categories in a multi-class problem or ranked preferences.

ISO Meaning, Origin and Examples • 7ESL
ISO Meaning, Origin and Examples • 7ESL
Did Vs Osdd, Osdd System, Osdd-1b System, Osdd Systems, Therapy Counseling, Dissociation, Mental Disorders, Mental And Emotional Health, I Dont Have Friends
Did Vs Osdd, Osdd System, Osdd-1b System, Osdd Systems, Therapy Counseling, Dissociation, Mental Disorders, Mental And Emotional Health, I Dont Have Friends
OSI Model
OSI Model
show me your essa’s in the comments!!
show me your essa’s in the comments!!
Vowel Teams OA & OE Interactive SeeSaw Lesson - OG Aligned
Vowel Teams OA & OE Interactive SeeSaw Lesson - OG Aligned
osmosis vs diffusion: key differences every biology student must know
osmosis vs diffusion: key differences every biology student must know

A typical example of a Closed-Set problem is image recognition. Feeding an image of an animal to a model, the model must choose its prediction from a pre-determined set of animal classes (cat, dog, bird, etc.). It doesn't need to handle unknown classes.

Closed-Set Problem Examples

Here are two more examples of Closed-Set problems:

AD
  1. Sentiment Analysis: Input text is classified into predefined categories such as positive, negative, or neutral.
  2. Multiple Choice Questions: The task is to select the correct answer from a given set of choices.

Lack of Flexibility in Closed-Set Problems

While Closed-Set problems are more straightforward to solve, they lack the adaptability of Open-Set problems. An algorithm trained on a specific set of categories will not perform well on new, unseen categories.

OSAS-COMP NOUN
OSAS-COMP NOUN

For example, in sentiment analysis, an algorithm trained only to recognize 'positive' and 'negative' sentiments might struggle with nuanced sentiments like 'somewhat positive' or miscategorize sarcasm as 'negative'.

The flexibility versus predictability trade-off between Open-Set and Closed-Set problems is a crucial consideration when designing and selecting machine learning models. Just as understanding the nature of the problem at hand is key to applying the right solution, so too with Open-Set and Closed-Set problems. It's not about which is better; it's about choosing the right tool for the right job.

AD
AD