Why is task2vec agnostic to the model?

[brando90]

I saw this in the paper:

TASK2VEC depends solely on the task, and ignores
interactions with the model,...

why is this true?

[brando90]

cross posted in SO: https://stats.stackexchange.com/questions/590763/why-is-the-embedding-of-a-task-using-task2vec-does-not-depend-on-the-model

[brando90]

https://stats.stackexchange.com/questions/590763/why-is-the-embedding-of-a-task-using-task2vec-does-not-depend-on-the-model

https://ai.stackexchange.com/questions/37261/why-is-the-embedding-of-a-task-using-task2vec-not-depend-on-the-model

[brando90]

https://stats.stackexchange.com/questions/590763/why-is-the-embedding-of-a-task-using-task2vec-does-not-depend-on-the-model

https://ai.stackexchange.com/questions/37261/why-is-the-embedding-of-a-task-using-task2vec-not-depend-on-the-model

https://www.reddit.com/r/learnmachinelearning/comments/xu9ycc/why_is_the_embedding_of_a_task_using_task2vec_not/

[brando90]

https://stats.stackexchange.com/questions/590763/why-is-the-embedding-of-a-task-using-task2vec-does-not-depend-on-the-model

https://ai.stackexchange.com/questions/37261/why-is-the-embedding-of-a-task-using-task2vec-not-depend-on-the-model

https://www.reddit.com/r/learnmachinelearning/comments/xu9ycc/why_is_the_embedding_of_a_task_using_task2vec_not/

https://www.reddit.com/r/MachineLearning/comments/xu9was/why_is_the_embedding_of_a_task_using_task2vec_not/

[brando90]

https://stats.stackexchange.com/questions/590763/why-is-the-embedding-of-a-task-using-task2vec-does-not-depend-on-the-model

https://ai.stackexchange.com/questions/37261/why-is-the-embedding-of-a-task-using-task2vec-not-depend-on-the-model

https://www.reddit.com/r/learnmachinelearning/comments/xu9ycc/why_is_the_embedding_of_a_task_using_task2vec_not/

https://www.reddit.com/r/MachineLearning/comments/xu9was/why_is_the_embedding_of_a_task_using_task2vec_not/

https://www.quora.com/unanswered/Why-is-the-embedding-of-a-task-using-Task2Vec-not-dependent-on-the-model

[mohd-faizy]

Task2vec is agnostic to the model because it does not rely on any specific machine-learning model. Instead, it uses a simple vector representation of data that can be used with any machine learning algorithm i.e it does not require any specific information about the model in order to create vector representations of the task. This makes it more flexible and applicable to a wider range of tasks and domains.

TASK2VEC depends solely on the task and ignores
interactions with the model,...

There are a few reasons why TASK2VEC might ignore interactions with the model. First, the task might be very simple and well-defined, such that the model is not needed to understand it.

Second, the task might be so complex that the model is not able to learn anything from it.

Finally, TASK2VEC might be designed to be used with a specific model, and therefore is not designed to work with other models.