Questions tagged [long-short-term-memory]

For questions related to the long-short term memory (LSTM), which refers to a recurrent neural network architecture that uses LSTM units. The first LSTM unit was proposed in 1997 by Sepp Hochreiter and Jürgen Schmidhuber in the paper "Long-Short Term Memory".

Filter by
Sorted by
Tagged with
28
votes
4answers
44k views

How to select number of hidden layers and number of memory cells in an LSTM?

I am trying to find some existing research on how to select the number of hidden layers and the size of these of an LSTM-based RNN. Is there an article where this problem is being investigated, i.e., ...
10
votes
1answer
7k views

How does LSTM in deep reinforcement learning differ from experience replay?

In the paper Deep Recurrent Q-Learning for Partially Observable MDPs, the author processed the Atari game frames with an LSTM layer at the end. My questions are: How does this method differ from the ...
10
votes
2answers
4k views

How to train a chatbot

I wanted to started experimenting with neural network and as a toy problem I wished to train one to chat, i.e. implement a chatting bot like cleverbot. Not that clever anyway. I looked around for ...
8
votes
2answers
660 views

Why use a recurrent neural network over a feedforward neural network for sequence prediction?

If recurrent neural networks (RNNs) are used to capture prior information, couldn't the same thing be achieved by a feedforward neural network (FFNN) or multi-layer perceptron (MLP) where the inputs ...
8
votes
1answer
129 views

How does the forget layer of an LSTM work?

Can someone explain the mathematical intuition behind the forget layer of an LSTM? So as far as I understand it, the cell state is essentially long term memory embedding (correct me if I'm wrong), ...
7
votes
2answers
1k views

Structure of LSTM RNNs

I have some very basic questions here. This is probably because I didn't read the relevant documents closely enough. If I used some terminology incorrectly, please point them out. Thank you! For ...
6
votes
2answers
8k views

Can LSTM Nets be speed up by GPU?

I am training LSTM Nets with Keras on a small mobile GPU. The speed on GPU is slower then on CPU. I found some articles that say that it is hard to train LSTMs (RNNs) on GPUs because the training ...
6
votes
2answers
157 views

Shortening the development time of a neural network

I am developing an LSTM for sequence tagging. During the development, I do various changes in the system, for example, add new features, change the number of nodes in the hidden layers, etc. After ...
6
votes
1answer
897 views

Will attention based networks prevail over RNN and LSTM?

There is no point in picking one of the growing number of articles that come up in a web search for, "Deep learning attention networks," however the bold claims in Attention Is All You Need, Ashish ...
6
votes
1answer
3k views

Where can I find the original paper that introduced RNNs?

I was able to find the original paper on LSTM, I was not able to find the paper that introduced "vanilla" RNNs. Where can I find it?
5
votes
4answers
351 views

Use Machine/Deep Learning to Guess a String

I want to be able to input a block of text and then have it guess a string within a predefined range (i.e. a string that starts with three letters and ends with five numbers like "XXX12345", etc). ...
4
votes
1answer
77 views

Can the decoder in a transformer model be parallelized like the encoder?

Can the decoder in a transformer model be parallelized like the encoder? As far as I understand, the encoder has all the tokens in the sequence to compute the self-attention scores. But for a decoder,...
4
votes
2answers
64 views

Why are GRU and LSTM better than standard RNNs?

It seems that older RNNs have a limitation for their use cases and have been outperformed by other recurrent architectures, such as the LSTM and GRU.
4
votes
1answer
130 views

How are LSTM's trained for text generation?

I've seen some articles about text generation using LSTMs (or GRUs) for text generation. Basically it seems you train them by folding them out, and putting a letter in each input. But say you trained ...
4
votes
2answers
36 views

Using a neural network to identify a stable region within a set of data?

I am working on a problem in which I am attempting to find a stable region in a spiral galaxy. The PI I'm working with asked me to use machine learning as a tool to solve the problem. I have created ...
4
votes
1answer
122 views

What would be the best approach to teach an AI to learn how to “sing” along a beat?

I have heard and read about HyperGAN, LSTM and a few other techniques, but I have a hard time piecing the overall concept together. End Goal Being able to input an instrumental and get an output of ...
4
votes
1answer
400 views

Over- and underestimations of the lowest and highest values in LSTM network

I'm training a LSTM network with multiple inputs and several LSTM layers in order to setup a time series gap filling procedure. The LSTM is trained bidirectionally with "tanh" activation on the ...
4
votes
0answers
49 views

RNN models displays upper limit on predictions

I have trained a RNN, GRU, and LSTM on the same dataset, and looking at their respective predictions I have observed, that they all display an upper limit on the value they can predict. I have ...
4
votes
1answer
28 views

Training an RNN to answer simple quesitons

I would like to train an RNN to follow the sentences: "Would you like some cheese"? with "Yes, I would like some cheese." So whenever the template "Would you like some ____?" appears then RNN ...
4
votes
1answer
130 views

dimensions of hidden layer and cell state layer in LSTM

I was following some examples to get familiar with tensorflow LSTM related api, but noticed that all LSTM initialization functions require only num_units parameter which denotes number of hidden units ...
3
votes
2answers
48 views

How to use LSTM to generate a paragraph

A LSTM model can be trained to generate text sequences by feeding the first word. After feeding the first word, the model will generate a sequence of words (a sentence). Feed the first word to get the ...
3
votes
1answer
47 views

Can dropout layers not influence LSTM training?

I am working on a project that requires time-series prediction (regression) and I use LSTM network with first 1D conv layer in Keras/TF-gpu as follows: ...
3
votes
1answer
29 views

Can non-sequential deep learning models outperform sequential models in time series forecasting?

Can a CNN (or other non-sequential deep learning models) outperform LSTM (or other sequential models) in time series data? I know this question is not very specific, but I experienced this when ...
3
votes
1answer
26 views

Structure discrepancy of an LSTM?

I've found multiple depictions of how an LSTM cell operates. See 2 below: and Each of these images suggest the hidden state is utilised differently. On the top diagram, it is shown that the hidden ...
3
votes
1answer
134 views

Object IN/OUT counting using CNN+RNN

I am building a video analytics program for counting moving things in a video. I am detecting bicycles and nothing else. I run object detection using the SSD mobile-net model in all the frames and ...
3
votes
2answers
118 views

Price Movement Forecasting Issue

I am working on a project for price movement forecasting and I am stuck with poor quality predictions. At every time-step I am using an LSTM to predict the next 10 time-steps. The input is the ...
3
votes
1answer
47 views

Using different timesteps for features and target value

I would like to know whether it's wrong; when working with time series data; to use daily prices as features and the price after 3 days as target. Is this correct or should I use the next-day price ...
3
votes
2answers
108 views

How should the output layer of an LSTM be when the output are word embeddings?

I'm having trouble grasping how to output word embeddings from an LSTM model. I'm seeing many examples using a softmax activation function on the output, but for that I would need to output one hot ...
3
votes
0answers
42 views

Why don't the neural networks inside LSTM cells contain hidden layers?

I watched a video explaining how LSTM cells have very rudimentary feed-forward neural networks, basically a 2 layer input-output with no hidden layers. Why don't LSTM cells have more complex neural ...
3
votes
0answers
27 views

How can I do hyperparameter optimization for a CNN-LSTM neural network?

I have built a CNN-LSTM neural network with 2 inputs and 2 outputs in Keras. I trained the network with model.fit_generator() (and not ...
3
votes
1answer
54 views

Why can't LSTMs keep track of the “important parts” of a sequence?

I keep reading about how LSTMs can't remember the "important parts" of a sequence which is why attention-based mechanisms are required. I was trying to use LSTMs to find people's name format. For ...
3
votes
0answers
78 views

Why do regression LSTMs learn high to low inputs significantly better than low to high?

The specific problem I have is learning the relation $x^2$. I have an array of 0 through 19 (input values) and a target array of 0, 1, 4, 9, 16, 25, 36 and so on all the way up to $19^2$=361. I have ...
3
votes
0answers
20 views

What evaluation metric are used for sequence-to-sequence prediction problems?

I am solving many sequence-to-sequence prediction problems using RNN/LSTM. What type of evaluation metrics can be used for sequence prediction problems? One metric is the mean squared error (MSE) ...
3
votes
1answer
49 views

What sort of Neural Network is best suited to predicting a future purchase?

I have previously implemented a Neural Network with Back-Propagation that was able to learn Tic-tac-toe and could go pretty well at Connect-4. Now I'm trying to do a NN that can make a prediction. ...
3
votes
0answers
120 views

What can be considered a deep recurrent neural network?

In the paper Deep Recurrent Q-Learning for Partially Observable MDPs, the DRQN is described as DQN with the first post-convolutional fully-connected layer replaced by a recurrent LSTM. I have DQN ...
3
votes
0answers
40 views

How do the relative number of cells between neighboring stacked LSTM layers affect the network's behavior?

It seems that stacking LSTM layers can be beneficial for some problem settings in order to learn higher levels of abstraction of temporal relationships in the data. There is already some discussion on ...
3
votes
1answer
1k views

Seq2Seq dialogs predicts only most common words like `you` after couple of epoches

I'm training Seq2Seq model on OpenSubtitles dialogs - Cornell-Movie-Dialogs-Corpus. My work based on the following papers (but currently I'm not implemented Attention yet): Sequence to Sequence ...
3
votes
0answers
128 views

Training RNN's on text: Can you use an ASCII encoding just as well as a one-hot character encoding?

I've mostly seen (e.g. in http://karpathy.github.io/2015/05/21/rnn-effectiveness/) that when training RNN's on text for something like language modeling, the text is usually featurized character-by-...
2
votes
1answer
1k views

What is the difference between ConvLSTM and CNN LSTM?

What will be the difference when used for video classification? Will they yield different results or are they the same fundamentally?
2
votes
1answer
68 views

Is my Neural Network program fully connected?

I have the following program for my neural network: ...
2
votes
1answer
207 views

What is a state in a recurrent neural network?

I am Reading "Supervised Sequence Labelling with Recurrent Neural Networks" written by Alex Graves to try to understand LSTM networks and I am a bit confused about the equations. Specifically, what I ...
2
votes
2answers
59 views

What is hidden state exactly in LSTM and RNN?

I'm working on research rn using LSTM as an encoder decoder in hopes to make inferences. The reason we are using encoder decoder for this is because there is hopes that the hidden state given by the ...
2
votes
1answer
2k views

Adding BERT embeddings in LSTM embedding layer

I am planning to use BERT embeddings in the LSTM embedding layer instead of the usual Word2vec/Glove Embeddings. What are the possible ways to do that?
2
votes
2answers
181 views

Why do we need multiple LSTM units in a layer?

What is the point of having multiple LSTM units in a single layer? Surely if we have a single unit it should be able to capture (remember) all the data anyway and using more units in the same layer ...
2
votes
1answer
438 views

Backward Pass for LSTMs

TL;DR I am currently trying to understand the mathematics in Ger's paper Long Short-Term Memory in Recurrent Neural Networks. I have found the document clear and readable so far. On pg. 21 of the ...
2
votes
0answers
21 views

How are batch statistics computed in Recurrent Batch Normalization?

I'm implementing recurrent BN per this paper in Keras, but looking at it and those citing it, a detail remains unclear to me: how are batch statistics computed? Authors omit explicit clarification, ...
2
votes
1answer
16 views

How to represent integer values in sequence to sequence prediction task in encoder-decoder LSTM?

I have a large 2D grid having 30k rows and 35k columns, so a total of 30x35k grid cells. Each grid cell is represented by a unique integer number (identity of grid cell). I have several trajectories ...
2
votes
0answers
37 views

How does a Bidirectional RNN work?

Could it be possible to reach a similar output via feeding a unidirectional network with the original data and the data played backwards?
2
votes
1answer
61 views

Why does an LSTM cycle on initialisation?

I initialised an LSTM with Xavier initialisation, although I've found this occurs for all initialisations I have tested. When initialised, if the LSTM is tested with a random input, it will get stuck ...
2
votes
0answers
19 views

How can I detect fast and slow motion in videos?

I'm trying to detect if a given video shot is fast or slow motion. Basically, I need to calculate a "video motion" score in a given video sequence, meaning how fast or slow motion the video is. For ...