Questions tagged [papers]
For questions related to artificial intelligence research papers. So, you should use this tag if you want someone to clarify something in a research paper.
331 questions
0
votes
0
answers
12
views
Does ARC-AGI 3 allow rendering JSON grids into images for multimodal LLMs?
There has been a lot of criticism of ARC-AGI 3 online. A common argument is that humans have a perceptual advantage since we observe a rendered grid, while LLMs only receive a JSON representation. ...
0
votes
1
answer
34
views
How do I make sense of the implementation of adaptive prefix tuning?
With reference to this paper Towards Adaptive Prefix Tuning for Parameter-Efficient Language Model Fine-tuning, I have following difficulties in understanding the implementation:
• "$h_{i−1}$ is ...
- 101
4
votes
1
answer
303
views
Hassija et al., "Interpreting Black-Box Models", 2024: Is this a serious paper?
"Cognitive Computation" appears to be a serious peer-reviewed Springer journal. This paper was published therein in 2024: Hassija, Vikas/Vinay Chamola/Atmesh Mahapatra/Abhinandan Singal/...
- 43
1
vote
1
answer
96
views
Token compression in native sparse attention
I have a question about the token compression in the native sparse attention in https://arxiv.org/pdf/2502.11089.
When we compute the attention of $q_t$ and $K^{\sim cmp
}_t$, is $K^{\sim cmp
}_t$ ...
- 121
1
vote
1
answer
165
views
What is the paper that introduced the settings used in transformers?
What paper introduced settings like system prompt, temperature, top P, top K, repetition penalty, etc. that are used in transformers? Was in the original Attention is All You Need paper?
user91131
1
vote
1
answer
213
views
Trust region policy optimization guarantees for continuous case
The original TRPO paper stated an algorithm that used optimization of the following surrogate objective:
$$
L_\pi(\tilde{\pi})=\eta(\pi)+\sum_s \rho_\pi(s) \sum_a \tilde{\pi}(a \mid s) A_\pi(s, a)
$$
...
- 131
2
votes
1
answer
227
views
Confusion between versions of trust-region policy optimziation
The original TRPO paper stated an algorithm that used optimization of the following surrogate objective:
$$
L_\pi(\tilde{\pi})=\eta(\pi)+\sum_s \rho_\pi(s) \sum_a \tilde{\pi}(a \mid s) A_\pi(s, a)
$$
...
- 131
1
vote
1
answer
217
views
Which factors do distinguish between human and LLMs written text at the moment?
I saw OpenAI discontinue their GPT detection tool even though is was performing fairly well according to this paper:
https://doi.org/10.48550/arXiv.2306.05524
Why to discontinue it, is it really ...
- 61
1
vote
2
answers
165
views
Why are ergodic MDPs also communicating?
An MDP is ergodic if the Markov chain induced by any policy is
ergodic, which means any state is reachable from any other state by
following a suitable policy. [Source]
The part after "which ...
- 273
0
votes
0
answers
18
views
Definition of Effect Identifiability in 'On Pearl's Hierarchy and Foundations of Causal Inference'
You can find the afore-mentioned text here. I'm asking a clarification of Definition 17, which is as follows:
Definition 17 (Effect Identifiability) The causal effect of an action $do(X=x)$ on a set ...
- 101
2
votes
1
answer
116
views
How exactly is the dynamical unfolding implemented in ByteNet?
I am thinking about making use of ByteNet (https://arxiv.org/abs/1610.10099) architecture for a project, and would like to get a better understanding of how the model works.
I've read through the ...
1
vote
1
answer
86
views
Why is this RL derivation right?
This comes from the paper, Reinforcement Learning and Control as Probabilistic Inference: Tutorial and Review.
I don't know the why the following derivation is true. The paper only briefly explains ...
- 87
0
votes
1
answer
48
views
How would I approach finding the source code for the paper that is being discussed in this article?
I have read a news article relating to recent research into graph theory-based fraud detection using machine learning, but cannot find the source code for the paper discussed. The paper can be found ...
2
votes
0
answers
134
views
Should intelligent machines have a body? (Reference request)
I was studying AI when a question came to my mind.
I have to write a 6-pages essay on a chosen topic in AI, and I think it can be interesting to reflect on the measure in which machines should have a ...
- 193
1
vote
2
answers
2k
views
What is the difference between RAG-Sequence Model and RAG-Token Model?
When I start to learn about RAG, I can't understand the difference between the RAG-Sequence Model and RAG-Token Model.
First of all, I see that for RAG-Sequence Model we use just one document to ...
