All Questions
11,189
questions
0
votes
0
answers
8
views
LLM-like architecture capable of dynamically learning from its own output
Language Learning Models (LLMs) have demonstrated remarkable capabilities in quick learning during inference. They can effectively grasp a concept from a single example and generate relevant outputs. ...
- 173
0
votes
0
answers
11
views
Meaning of roles in the API of GPT-4/ChatGPT (system/user/assistant)
In the API of GPT-4 and ChatGPT, the prompt for a chat conversation is a list of messages, each marked as one of three roles: system, ...
- 272
0
votes
0
answers
5
views
Voice identification model
I'd like to know if there are any pre-trained TensorFlow/PyTorch models or guidance for voice identification in audio files. I want to use some versions of my own voice as input and identify all ...
- 101
3
votes
0
answers
14
views
Why does training converges when the norm of gradient increases?
This is from deep learning book by Ian Goodfellow and Yoshua Bengio and Aaron Courville.
When training converges well, I thought the gradient should be at local minima. But the book says it often does ...
- 143
0
votes
0
answers
10
views
What else beyond convnets is needed to detect sameness in a Bongard problem?
In her book Artificial Intelligence - A Guide for Thinking Humans, Melanie Mitchell explains that convolutional neural networks fail to build a useful model for these two classes of images, yielding ...
0
votes
0
answers
17
views
Detecting deepfake video
I want to create a tool for deepfake detection, that takes the input as audio clip, image or video clip and tell which part of it is fake.
Is there any convenient way to start developing this ...
- 171
0
votes
0
answers
17
views
Can I reduce computation by only predicting response tokens in a transformer and still get the same gradients?
I have been looking at the source code of the Stanford Alpaca model and I believe that during inference, the whole instruction + response data is fed into the model normally. Then the instruction part ...
0
votes
0
answers
10
views
Is it possible to apply entity fixed effects to a firth logit test?
All,
Currently working with a large dataset (~1.6 million observations) with a relatively low number of rare events (~600ish). To summarize, I'm looking at the impact of 7 foreign-policy related ...
0
votes
0
answers
23
views
How does a LLM (specifically transformer) pick words from it's vocabulary?
I have a very rough understanding of the "attention/self attention" mechanism of transformer models and how this can be used to process a set of word vectors provided as an input/prompt to ...
0
votes
0
answers
13
views
Why is my loss graph heavily fluctatuing?
I am working on a CNN project on an image dataset. I am applying Early Stopping technique in order to train the model. However, after training the model and obtaining the loss graph, it is heavily ...
- 1
0
votes
0
answers
12
views
Is manual binding output to input also an AI?
I know AI is primarly training a machine by samples of input-output in order it would learn itself about relations between the input and the output.
What if I manually add the relations? Is that still ...
- 1
0
votes
0
answers
3
views
Filter in the Single Shot Detector
Let's say I want to implement a single shot detector. When I get a feature mal as an input, I will use a 3x3 filter for prediction for each cell.
Let's say we have 5 classes with 6 ancors, I would ...
0
votes
0
answers
10
views
Transfer the colors of a face from a photo inside a drawing
I have a photo of a real person.
I have a colorized picture of a resembling person (same angle, different lighting and colors, drawing look alike)
I would like to transfer the colors of the photo ...
- 1
0
votes
0
answers
8
views
Looking for scalable spatiotemporal clustering algorithm or scalable mixed features data clustering algorithms
I want to cluster the Porto taxi dataset from https://www.kaggle.com/competitions/pkdd-15-taxi-trip-time-prediction-ii/data . After processing the dataset, I have a dataset which consists of temporal ...
1
vote
0
answers
26
views
What's the most efficient way of performing batched training of Causal Language Models?
I have seen a number of ways to train (yes, train, not fine-tune) these models efficiently with batches. I will illustrate these techniques with the following example dataset and context window:
...
- 271
0
votes
0
answers
33
views
Using reinforcement learning with an undefined set of actions and observations
I am new to RL and want to build a tool that will make certain decisions in order to navigate through a web page to which I own the code. My goal is the agent to take different actions based on the ...
- 101
0
votes
0
answers
14
views
How can a neural network handle different image sizes, when performing deep dreaming using image pyramids?
I read that image pyramids are used for deep dreaming. The image is feed through the network multiple times and is rescaled during the process, to capture features in different scales. But how can the ...
- 1
1
vote
1
answer
77
views
What temperature would you recommend for the chatgpt api?
I believe that it is recommended to have a tiny bit of temperature with GPT 3 even for noncreative tasks like 0.2 or something (I am not entirely sure why).
Last I checked, and if I remember correctly,...
- 111
1
vote
0
answers
13
views
Best calculus books for Deep Learning
Recommend some calculus books for Deep Learning and neural networks. I know what is integration, differentiation, derivates, limits on a based level. I would like to understand on deep level the ...
- 11
2
votes
2
answers
198
views
Is the output of image generation models like Midjourney and Stable Diffusion deterministic?
Assuming the user can set all parameters, including but not limited to the seed.
Is the output deterministic? As in, the same set of inputs will create the same image?
0
votes
0
answers
10
views
Pedestrian/Object detection
Today automobiles have many kinds of detection systems and I'm currently researching one on Pedestrian Detection systems.
I haven't figured exactly out the difference between these three systems. If ...
- 1
0
votes
1
answer
16
views
Implementing deep learingn paper from scratch
I am fairly new to deep learning. I have understood theory part of the subject.
My knowledge of python is decent.
As a project I want to implement a deep learning paper from scratch (does not have to ...
0
votes
1
answer
22
views
Q learning: Action is combination of two actions
Suppose in my example I want an agent to learn a behavior that is made up of a combination of actions.
So consider the following example with a tamagotchi like game: There are 5 pets and 3 actions ...
- 101
1
vote
1
answer
29
views
Does MSE loss function work in NN training for predicting values between 0-1?
In a NN regression problem, considering that MSE is squaring the error and the error is between 0 and 1 would it be pointless to use MSE as our loss function during model training?
For example:
...
0
votes
0
answers
7
views
In the proof of Neural Tangent Kernel stays constant in infinite width limit, why the norm of the dual mapping operator equals operator norm of kernel
For a fixed distribution $p^{in}$ on the input space $ \mathbb{R}^{n_0}$,
consider a function space $\mathcal{F}$ defined as $\{{f: \mathbb{R}^{n_0} \rightarrow \mathbb{R}^{n_L}}\}$.
On this space, ...
0
votes
1
answer
29
views
Recognize patterns within random sequences
I am familiar with ANNs as I studied them back in the days for regression and currently I'm working with CNN's for image recognition. But recently I was reading more about pattern recognition in ...
- 101
0
votes
0
answers
19
views
Role of $f$ Target Network in DDPG
I am trying to create a variant of DDPG in MATLAB that has no action-value $\langle Q \rangle$ net, but that instead works with networks $\langle V \rangle, \langle f \rangle, \langle r \rangle$, and ...
0
votes
0
answers
13
views
Clustering by number of spacial nodes in an area
I have a huge a amount of coordinates (hundred of millions) and I want to group them in areas, where every area should have approximately the same number of nodes and no more than a limit.
I was ...
- 101
0
votes
0
answers
11
views
In which process does filter and watermark take place?
In LLM, in order to avoid discrimination and abuse, the author will add filter and watermark functionalities. Could you tell me in which process these functionalities take place? Is it in weights(...
- 133
2
votes
0
answers
23
views
How can i tinker my neural network to learn stronger on rare events?
I am training a neural network on a regression problem.
Most of the time the actual y (label) has the same value (say ~0.2) and only in rare cases the actual y is very different (say 2.0 or -2.0)
...
- 21
0
votes
1
answer
18
views
Does changing deep learning model loss function affect the occurence of plateau?
If we use two loss functions that are in the same "units" (say MAE and RMSE). Should training the same model with these two different loss functions result in plateau happening at a ...
0
votes
0
answers
8
views
How to calculate the loss for object tracking model
I am using a CNN model to detect the location of a tennis ball's center (not bounding box). the output is a 2d tensor after a softmax layer so each cell represent the probebility that this is where ...
0
votes
1
answer
28
views
Adding a Transposition Table to Monte Carlo Tree Search
I think I'm having a bit of trouble wrapping my head around how a transposition table functions:
As I understand it you can store a value (simulation result?) for a given game state in this (hash) ...
- 119
-1
votes
1
answer
38
views
I'd like some suggestions on how to effectively master machine learning. I'm new to this and I'm kinda lost [closed]
Thing is, I got a lot of resources to learn from and I have no idea which ones will be the most effective, and where to start. I've come across 2 kinds of books, the ones which talk about the ...
- 1
1
vote
0
answers
16
views
Can a language model detect that a concept assumed by humanity as true is false? [duplicate]
With the arrival of AI-based search engines, I wonder if these models, being trained on human texts, always offer the solution most accepted by most of humans.
This raises the question of whether ...
user69389
0
votes
0
answers
9
views
Which architecture allows a neural network to make external queries?
What architecture can be used to allow a neural network to make external queries?
I am thinking of Bing Chat's ability to perform web searches, but we could also consider, for example, or retrieving ...
- 659
0
votes
0
answers
13
views
Bellmann Error with Time Step
I define the return $G$ as the discrete time integral of the reward $G = \Delta t \sum_{t = 0, ..., T}r_t$. In supposing $\Delta t = 1$ throughout, my instructor wrote the following formula for the ...
0
votes
0
answers
6
views
How to get optimal Scaling with raw PyTorch+DDP?
I'm trying to install a distributed training environment on a compute cluster that I have. I happen to know from previous experience that often scaling up the batch size "naively" isn't very ...
- 306
-1
votes
0
answers
18
views
Problem with my DQN
I have a problem with my code. My AI couldn't make a good q_values expectations. I have no idea where is the problem.
'''
class Morpion : #Code du Jeu Morpion
def init(self) -> None:
self....
- 1
0
votes
0
answers
25
views
How to use "states" as vectors in Q-Learning?
How do I use states as vectors in Q-Learning or any other RL Algorithms. Let's say I have state as a vector with probabilities[0,1], and I have to take an action if the state is valid (with ...
0
votes
1
answer
58
views
How to approach a toy classification problem using a neural network?
The toy problem:
50 unique numbers are randomly selected from number 0 to 99.
If number 1 appears in the selection but number 2 doesn't, the selection is labelled as "1".
If number 2 ...
- 1
1
vote
3
answers
56
views
Where would I start if I wanted to create an AI agent to play a 2d game?
I am keen on creating a little project that can play a fairly basic 2D game (more complex than say, snake but not as complex as mario kart) and would like some pointers on where to begin. I'm entirely ...
- 11
0
votes
0
answers
21
views
Can I use this method to improve the VAE model?
My VAE model is not learning well.
The model's learning history shows that the reconstruction loss is as large as approximately 8000 or more, and the KL-Divergence loss is diverging, starting with an ...
- 17
0
votes
0
answers
18
views
Using CLIP embeddings with VisualBERT
I am trying to use CLIP embeddings to train a VisualBERT model for multimodal image classification. Following are the steps that I perform:
Generate embeddings from CLIP model for text(j1) and images(...
0
votes
0
answers
11
views
LRP (Layer wise relevance propagation ) backward pass for two layer LSTM Networks
I am trying to calculate relevance scores for each time step of the input of my network which is a time series of shape (batch_size = 7000,sequence_length = 20,number of features = 1).
my network has ...
- 1
0
votes
0
answers
25
views
As a user, can I fine-tune end-to-end with GPT-3 if I want to add custom layers?
I am looking at the OpenAI API, and it seems that the only way to get embeddings for a specific text is to use the embedding API. In other words, I can only use the embeddings as fixed features in my ...
- 111
0
votes
1
answer
30
views
What are possible reasons for the validation loss increasing with more data?
I trained a neural network on an NLP problem and compared the loss and BLEU score on the validation data with the same training parameters in two scenarios: a) when I trained on 25% of the data, b) ...
0
votes
1
answer
20
views
How to convert my test data in the same dimensionality as my train data
I have trained a VAE with jpg images. My latent space dimension has 768 features and when plotting the latent space it looks like this:
However, when I use the scikit learn tool LDA (Linear ...
- 3
0
votes
0
answers
14
views
What metrics are there to measure the similarity of two disjoint sets of points?
I'm training an AI model to predict where to place train stations along a train track.
Specifically, I want to feed my model some information about the train track. The model generates a series of ...
- 101
-1
votes
0
answers
24
views
What are the key challenges and limitations of using AI models like GPT-4 for social media management? [closed]
I am currently running an experiment to explore the potential of AI models, specifically ChatGPT based on GPT-4 architecture, in managing social media content. As part of the experiment, ChatGPT ...
