Alternative to back-propagation: Are there other popular alternatives to back-propagation through which a neural network can learn ?

More Titas De's questions See All

Support Vector Machines: How do I theoretically calculate the maximum number of support vectors required for a machine learning problem ?

For a machine learning problem, how do I theoretically calculate the maximum number of support vectors required to solve the machine learning problem ?

11 December 2019 6,932 8 View

Text Summarization: Are there standard NLP ( deep-learning ) techniques on paragraph summarization ?

Given a paragraph, are there standard techniques on paragraph summarization, like extracting relevant keywords. Anything in this direction will be help-ful

11 December 2019 2,771 6 View

Support Vector Machines: How to find the minimum number of support vectors for a machine learning classification problem ?

How do I theoretically calculate the minimum number of support vectors required to solve a machine learning problem ?

11 December 2019 9,368 7 View

Deep Learning - transfer learning from one image classification dataset to another - Scientific process ?

Hi, is there any quantifiable scientific process to understand whether a deep learning model trained on a certain image classification dataset can be used for transfer learning for image...

10 November 2019 5,722 2 View

Machine Learning: Given a classification dataset, how can we quickly figure out that our dataset is really noisy ?

How can we figure out that our dataset is really noisy - both in terms of the data-points and labels ?

10 November 2019 9,155 1 View

Machine Learning Classification: How can we derive to the (theoretical) minimum classification error ?

Given a dataset, is there a systematic way to come to a maximum classification accuracy, beyond which no algorithm or classifier can improve the accuracy ?

10 November 2019 8,015 7 View

Machine Learning: Say we are given a classification dataset. Is there any systematic way to reduce the noise in the training labels ?

Reduction of noise in training labels.

10 November 2019 6,079 3 View

Apart from Self-Organizing maps, what are the most common neural network techniques used for clustering ?

Clustering - Neural Netowrks

08 September 2018 4,900 4 View

A 10-layer ResNet CNN trained from scratch gives lower classification accuracy than the same without residual connections. Any possible reasons ?

CNNs, Deep Learning, ResNet, Residual Connections

06 July 2018 722 5 View

Regarding CNNs in deep learning, can anyone tell me in which cases transfer learning increases the error or decreases the accuracy ?

Deep Learning - Transfer Learning

06 July 2018 5,100 5 View

Feedback defines the constitution of an organism?

“Here is a thought experiment. Let's place Rodolpho Llinas's jarred-brain on top of a body (Fig. 1). I bet Llinas would argue that his jarred-brain retains its own consciousness, and the android...

11 August 2024 2,483 1 View

Self-Organizing Superorganisms—as envisaged by Nenad Sestan (2018)?

The rate of glucose consumption by the neocortex is reduced by over 80% during anesthesia (Sibson et al. 1998), which disables the synapses (Richards 2002) that are inundated by glial tissue (Engl...

08 August 2024 3,118 0 View

Measuring the Intelligence of a Species?

Larger brains, which typically contain more neurons, store and transfer more information (Tehovnik and Chen 2015), but the precise relationship between number of neurons and information has yet to...

05 August 2024 1,238 2 View

How can i do multivariate Time Series forecast using MLP, ANFIS and LSTM?

I need the python code to forecast what crop production will be in the next decade considering climate and crop production variables as seen in the attached.csv file.

05 August 2024 2,977 3 View

The Curse of Evolution and Complexity?

Brain and body mass together are positively correlated with lifespan (Hofman 1993). The duration of neural development is one of the best predictors of brain size, and conception is the best...

05 August 2024 6,247 3 View

Need help with my research project on open source SIEM and machine learning?

Hello everyone, I am currently working on a research project that aims to integrate machine learning techniques into an open source SIEM tool to automate the creation of security use cases from...

04 August 2024 3,196 2 View

Swimming/space travel depends on the proprioceptive muscle spindles?

When the entire neocortex is ablated in rodents, although they are still able to swim, all the limbs move continuously and asynchronously (Vanderwolf 2006; Vanderwolf et al. 1978). Normal animals...

03 August 2024 835 3 View

What are the limitations and challenges of using machine learning for predicting concrete compressive strength in practical applications?

Machine learning (ML) has shown great potential in predicting the compressive strength of concrete, an important property for structural engineering. However, its practical application comes with...

03 August 2024 2,546 2 View

Some new emerging problems on application of RL for scheduling in IoT networks?

I have seen plenty of existing works on applied Reinforcement Learning (RL) policies for optimized scheduling in IoT networks including Q-learning, DQNs, and the newer ones including PPO for...

01 August 2024 8,754 2 View

How to Compress Information Neurally?

Samuel Morse, the inventor of the Morse Code, understood that certain letters in the English language occurred more frequently than others (Gallistel and King 2010). To deal with this, Morse used...

01 August 2024 4,456 2 View

Fabrice Clerot

many pointers can be found here :

http://www.breloff.com/no-backprop/

http://www.breloff.com/no-backprop-part2/

and there always is the good old Hebb rule

https://www.ece.ucsb.edu/wcsl/people/aseem/Aseem_stuff/hebbian_preprint.pdf

Article Deep Learning With Asymmetric Connections and Hebbian Update...

also regarding BP through time

Preprint Biologically inspired alternatives to backpropagation throug...

now; i would not say these are "poipular" alternatives ... but these are definitely alternatives to BP learning.

Adnan Riaz

these are some alternatives to BP learning.

Shahbaa I. Khaleel

The Alternative to backpropagation through which a neural network can learn is the Elman neural network and Jordan neural network.

also there is many of learning rule to training neural network for example Hebbian learning , etc.

the attached file is contain many of learning rule.

Good luck

Samy Azer

Titas De-

Thank you for your excellent question. You may use genetic algorithm or other approaches- see the link below for abstracts-

https://www.sciencedirect.com/topics/engineering/genetic-algorithm

Titas De

Thank you Fabrice Clerot , Adnan Riaz , Shahbaa I. Khaleel and Samy Azer

Olawale Onabola

You may also check this paper from Google on approximate backprop

https://arxiv.org/abs/2001.06232