Artificial Intelligence Stack Exchange Community Digest

Top new questions this week:

Are softmax outputs of classifiers true probabilities?

BACKGROUND: The softmax function is the most common choice for an activation function for the last dense layer of a multiclass neural network classifier. The outputs of the softmax function have …

activation-functions probability softmax probability-theory

asked by Snehal Patel Score of 6 answered by D.W. Score of 4

How are gradients of individual layers computed?

I have been reading some papers recently (example: arxiv.org/pdf/2012.00363.pdf) which seem to be training individual layers of, say, a transformer, holding the rest of the model frozen/…

neural-networks natural-language-processing gradient-descent transformer

asked by nlp4892 Score of 2 answered by Dr. Snoopy Score of 1

What would be a good optimization technique for this kind of problem?

Problem Description: Since I am not sure if there is a scientific term that categorizes this problem, I will do my best to describe it thoroughly. Suppose there is a chamber that’s being filled with …

machine-learning optimization constrained-optimization particle-swarm-optimization

asked by Sobhan Score of 1

How would one go about checking numerical stability of the deep learning algorithms?

Let’s say I have a trained deep learning model. It would be good if it would be numerically stable, so if I change input by small amount, the output will also change by small amount. How should I …

deep-learning stability

asked by SlimakSlimak Score of 1 answered by Snehal Patel Score of 1

Interpretation of the roc curve on test set

Hi i got the following roc curve: What does this mean? has this to do with overfitting? Is my data wrong preprocessed? i do not understand and would appreciate an answer.

machine-learning

asked by ProgrammingBot Score of 1 answered by Snehal Patel Score of 1

Reinforcement Learning: sampled actions in k-armed bandit problem

I am reading the book Reinforcement Learning: An Introduction. Second edition (Richard S. Sutton and Andrew G. Barto). In the k-armed bandit problem using $\varepsilon$-greedy selection method, the …

reinforcement-learning multi-armed-bandits

asked by flamingheart Score of 1 answered by Neil Slater Score of 1

Solving Bin Assignment problems

What are some non-RL-based approaches to solving a typical bin assignment problem, i.e., given a set of items (can be multi-dimensional), find the bin/knapsack/target which best packs (with minimum …

machine-learning deep-learning classification

asked by helloworld Score of 1

Greatest hits from previous weeks:

Is AI living or non-living?

I’m a bit confused about the definition of life. Can AI systems be called ‘living’? Because they can do most of the things that we can. They can even communicate with one another. They are not …

philosophy definitions biology

asked by Soham Score of 12 answered by Lorry Laurence mcLarry Score of 7

What is the Bellman operator in reinforcement learning?

In mathematics, the word operator can refer to several distinct but related concepts. An operator can be defined as a function between two vector spaces, it can be defined as a function where the …

reinforcement-learning terminology math bellman-equations bellman-operators

asked by nbro Score of 24 answered by Philip Raeisghasem Score of 27

What are the differences between transfer learning and meta learning?

What are the differences between meta-learning and transfer learning? I have read 2 articles on Quora and TowardDataScience. Meta learning is a part of machine learning theory in which some …

machine-learning deep-learning comparison transfer-learning meta-learning

asked by Long Score of 7 answered by nbro Score of 5

Do scientists know what is happening inside artificial neural networks?

Do scientists or research experts know from the kitchen what is happening inside complex “deep” neural network with at least millions of connections firing at an instant? Do they understand …

neural-networks deep-learning convolutional-neural-networks explainable-ai

asked by kenorb Score of 94 answered by Maxim Score of 69

Could a neural network detect primes?

I am not looking for an efficient way to find primes (which of course is a solved problem). This is more of a “what if” question. So, in theory, could you train a neural network to predict …

neural-networks prediction primality-test

asked by Fullk33 Score of 37 answered by Douglas Daseeco Score of 17

What are “bottlenecks” in neural networks?

What are “bottlenecks” in the context of neural networks? This term is mentioned, for example, in this TensorFlow article, which also uses the term “bottleneck values”. How does …

neural-networks deep-learning tensorflow terminology bottlenecks

asked by Anurag Singh Score of 27 answered by Neil Slater Score of 36

What is the difference between “mutation” and “crossover”?

In the context of evolutionary computation, in particular genetic algorithms, there are two stochastic operations “mutation” and “crossover”. What are the differences between them?

comparison evolutionary-algorithms crossover-operators mutation-operators genetic-operators

asked by Abbas Ali Score of 5 answered by nbro Score of 5

Can you answer these questions?

Would the optimal policy remain same, if I replace R with V*?

In the context of RL, say I’m performing Value Iteration on a reward function R1. And the converged optimal policy is P1 and values are V1. Then, let’s say I set rewards to be R2=V1 and perform value …

reinforcement-learning markov-decision-process reward-functions value-iteration policy-iteration

asked by famishedrover Score of 1

Trying to understand where exactly is permutation happening in summation equation

I am reading an article of creating an embedding which is order-invariant to inputs ($m_{l}$). They refered to order-invariance in eq (5) as follows: The order-invariant operation happens in (5), …

long-short-term-memory attention

asked by Oculu Score of 1

How can (pretrained) language models actively seek additional training data – possibly reference request?

I am reading the paper “Large Language Models Can Self-Improve” arxiv.org/abs/2210.11610 in which the authors consider that LLM can generate Chain-of-Thoughts sequences and even …

language-model fine-tuning self-supervised-learning

asked by TomR Score of 1