# Optimization in deep learning- Learn with examples

June 24, 2022

Deep learning relies on optimization methods. Training a complicated deep learning model, on the other hand, might take hours, days, or even weeks. The training efficiency of the model is directly influenced by the optimization algorithm's performance. Understanding the fundamentals of different optimization algorithms and the function of their hyperparameters, on the other hand, will allow us to modify hyperparameters in a targeted manner to improve deep learning model performance.

In this blog, we'll go through some of the most popular deep learning optimization techniques in detail.

Table of Content:

1. The goal of Optimization in Deep learning
1. Gradient Descent Deep Learning Optimizer
1. Stochastic Gradient Descent Deep Learning Optimizer
1. RMSprop (Root Mean Square) Optimizer

## The goal of Optimization in Deep learning-

Although optimization may help deep learning by lowering the loss function, the aims of optimization and deep learning are fundamentally different. The former is more focused on minimizing an objective, whereas the latter is more concerned with finding a good model given a finite quantity of data. Training error and generalization error, for example, vary in that the optimization algorithm's objective function is usually a loss function based on the training dataset, and the purpose of optimization is to minimize training error. Deep learning (or, to put it another way, statistical inference) aims to decrease generalization error. In order to achieve the latter, we must be aware of overfitting as well as use the optimization procedure to lower the training error.

## Gradient Descent Deep Learning Optimizer-

Gradient Descent is the most common optimizer in the class. Calculus is used in this optimization process to make consistent changes to the parameters and reach the local minimum. Before you go any further, you might be wondering what a gradient is?

Consider that you are holding a ball that is lying on the rim of a bowl. When you lose the ball, it travels in the steepest direction until it reaches the bowl's bottom. A gradient directs the ball in the steepest way possible to the local minimum, which is the bowl's bottom.

Gradient descent works with a set of coefficients, calculates their cost, and looks for a cost value that is lower than the current one. It shifts to a lesser weight and updates the values of the coefficients. The procedure continues until the local minimum is found. A local minimum is a point beyond which it is impossible to go any farther.

For the most part, gradient descent is the best option. It does, however, have significant drawbacks. Calculating the gradients is time-consuming when the data is large. For convex functions, gradient descent works well, but it doesn't know how far to travel down the gradient for nonconvex functions.Gradient descent works well for convex functions

## Stochastic Gradient Descent Deep Learning Optimizer-

On large datasets, gradient descent may not be the best solution. We use stochastic gradient descent to solve the problem. The word stochastic refers to the algorithm's underlying unpredictability. Instead of using the entire dataset for each iteration, we use a random selection of data batches in stochastic gradient descent. As a result, we only sample a small portion of the dataset. The first step in this technique is to choose the starting parameters and learning rate. Then, in each iteration, mix the data at random to get an estimated minimum. When compared to the gradient descent approach, the path taken by the algorithm is full of noise since we are not using the entire dataset but only chunks of it for each iteration.

As a result, SGD requires more iterations to attain the local minimum. The overall computing time increases as the number of iterations increases. However, even when the number of iterations is increased, the computation cost remains lower than that of the gradient descent optimizer. As a result, if the data is large and the processing time is a consideration, stochastic gradient descent should be favored over batch gradient descent.

Mini batch SGD straddles the two preceding concepts, incorporating the best of both worlds. It takes training samples at random from the entire dataset (the so-called mini-batch) and computes gradients just from these. By sampling only a fraction of the data, it aims to approach Batch Gradient Descent.

We require fewer rounds because we're utilizing a chunk of data rather than the entire dataset. As a result, the mini-batch gradient descent technique outperforms both stochastic and batch gradient descent algorithms. This approach is more efficient and reliable than previous gradient descent variations. Because the method employs batching, all of the training data does not need to be placed into memory, making the process more efficient. In addition, the cost function in mini-batch gradient descent is noisier than that in batch gradient descent but smoother than that in stochastic gradient descent. Mini-batch gradient descent is therefore excellent and delivers a nice mix of speed and precision.

Mini-batch SGD is the most often utilized version in practice since it is both computationally inexpensive and produces more stable convergence.

Adagrad keeps a running total of the squares of the gradient in each dimension, and we adjust the learning rate depending on that total in each update. As a result, each parameter has a variable learning rate (or an adaptive learning rate). Furthermore, when we use the root of the squared gradients, we only consider the magnitude of the gradients, not the sign. We can observe that the learning rate is reduced when the gradient changes rapidly. The learning rate will be higher when the gradient changes slowly. Due to the monotonic growth of the running squared sum, one of Adagrad's major flaws is that the learning rate decreases with time.

## RMSprop (Root Mean Square) Optimizer-

Among deep learning aficionados, the RMS prop is a popular optimizer. This might be due to the fact that it hasn't been published but is nonetheless well-known in the community. RMS prop is a natural extension of RPPROP's work. The problem of fluctuating gradients is solved by RPPROP. The issue with the gradients is that some were modest while others may be rather large. As a result, establishing a single learning rate may not be the ideal option. RPPROP adjusts the step size for each weight based on the sign of the gradient. The two gradients are initially compared for signs in this technique.

To update network weights during training, this optimization approach is a further development of stochastic gradient descent. Unlike SGD, Adam optimizer modifies the learning rate for each network weight independently, rather than keeping a single learning rate for the entire training. The Adam optimizers inherit both Adagrad and RMS prop algorithm characteristics. Instead of using the first moment (mean) like in RMS Prop, Adam employs the second moment of the gradients to modify learning rates. We take the second instance of the gradients to imply the uncentered variance (we don't remove the mean).

AdaDelta is a more powerful variant of the AdaGrad optimizer. It is based on adaptive learning and is intended to address the major shortcomings of AdaGrad and the RMS prop optimizer. The fundamental disadvantage of the two optimizers mentioned above is that the starting learning rate must be set manually. Another issue is the decreasing learning rate, which eventually becomes infinitesimally tiny. As a result, after a given number of iterations, the model can no longer acquire new information.

## Conclusion-

This is a comprehensive explanation of the various optimization methods utilized in Deep Learning. We went through three different types of gradient descent and then moved on to additional optimizer techniques. There is still a lot of work to be done in the field of optimization.

However, for the time being, it is critical to understand your needs and the type of data you are working with in order to select the finest optimization technique and obtain excellent outcomes.

Latest Blogs
October 18, 2022

### A Complete Guide To Customer Acquisition For Startups

Any business is enlivened by its customers. Therefore, a strategy to constantly bring in new clients is an ongoing requirement. In this regard, having a proper customer acquisition strategy can be of great importance.

The problem with customer acquisition

As an organization, when working in a diverse and competitive market like India, you need to have a well-defined customer acquisition strategy to attain success. However, this is where most startups struggle. Now, you may have a great product or service, but if you are not in the right place targeting the right demographic, you are not likely to get the results you want.

To resolve this, typically, companies invest, but if that is not channelized properly, it will be futile.

So, the best way out of this dilemma is to have a clear customer acquisition strategy in place.

How can you create the ideal customer acquisition strategy for your business?

• Define what your goals are

You need to define your goals so that you can meet the revenue expectations you have for the current fiscal year. You need to find a value for the metrics –

• MRR – Monthly recurring revenue, which tells you all the income that can be generated from all your income channels.
• CLV – Customer lifetime value tells you how much a customer is willing to spend on your business during your mutual relationship duration.
• CAC – Customer acquisition costs, which tells how much your organization needs to spend to acquire customers constantly.
• Churn rate – It tells you the rate at which customers stop doing business.

All these metrics tell you how well you will be able to grow your business and revenue.

You need to understand who your current customers are and who your target customers are. Once you are aware of your customer base, you can focus your energies in that direction and get the maximum sale of your products or services. You can also understand what your customers require through various analytics and markers and address them to leverage your products/services towards them.

• Choose your channels for customer acquisition

How will you acquire customers who will eventually tell at what scale and at what rate you need to expand your business? You could market and sell your products on social media channels like Instagram, Facebook and YouTube, or invest in paid marketing like Google Ads. You need to develop a unique strategy for each of these channels.

If you know exactly what your customers have in mind, then you will be able to develop your customer strategy with a clear perspective in mind. You can do it through surveys or customer opinion forms, email contact forms, blog posts and social media posts. After that, you just need to measure the analytics, clearly understand the insights, and improve your strategy accordingly.

Combining these strategies with your long-term business plan will bring results. However, there will be challenges on the way, where you need to adapt as per the requirements to make the most of it. At the same time, introducing new technologies like AI and ML can also solve such issues easily. To learn more about the use of AI and ML and how they are transforming businesses, keep referring to the blog section of E2E Networks.

https://www.helpscout.com/customer-acquisition/

https://www.cloudways.com/blog/customer-acquisition-strategy-for-startups/

https://blog.hubspot.com/service/customer-acquisition

October 18, 2022

### Image-based 3D Object Reconstruction State-of-the-Art and trends in the Deep Learning Era

3D reconstruction is one of the most complex issues of deep learning systems. There have been multiple types of research in this field, and almost everything has been tried on it — computer vision, computer graphics and machine learning, but to no avail. However, that has resulted in CNN or convolutional neural networks foraying into this field, which has yielded some success.

## The Main Objective of the 3D Object Reconstruction

Developing this deep learning technology aims to infer the shape of 3D objects from 2D images. So, to conduct the experiment, you need the following:

• Highly calibrated cameras that take a photograph of the image from various angles.
• Large training datasets can predict the geometry of the object whose 3D image reconstruction needs to be done. These datasets can be collected from a database of images, or they can be collected and sampled from a video.

By using the apparatus and datasets, you will be able to proceed with the 3D reconstruction from 2D datasets.

State-of-the-art Technology Used by the Datasets for the Reconstruction of 3D Objects

The technology used for this purpose needs to stick to the following parameters:

• Input

Training with the help of one or multiple RGB images, where the segmentation of the 3D ground truth needs to be done. It could be one image, multiple images or even a video stream.

The testing will also be done on the same parameters, which will also help to create a uniform, cluttered background, or both.

• Output

The volumetric output will be done in both high and low resolution, and the surface output will be generated through parameterisation, template deformation and point cloud. Moreover, the direct and intermediate outputs will be calculated this way.

• Network architecture used

The architecture used in training is 3D-VAE-GAN, which has an encoder and a decoder, with TL-Net and conditional GAN. At the same time, the testing architecture is 3D-VAE, which has an encoder and a decoder.

• Training used

The degree of supervision used in 2D vs 3D supervision, weak supervision along with loss functions have to be included in this system. The training procedure is adversarial training with joint 2D and 3D embeddings. Also, the network architecture is extremely important for the speed and processing quality of the output images.

• Practical applications and use cases

Volumetric representations and surface representations can do the reconstruction. Powerful computer systems need to be used for reconstruction.

Given below are some of the places where 3D Object Reconstruction Deep Learning Systems are used:

• 3D reconstruction technology can be used in the Police Department for drawing the faces of criminals whose images have been procured from a crime site where their faces are not completely revealed.
• It can be used for re-modelling ruins at ancient architectural sites. The rubble or the debris stubs of structures can be used to recreate the entire building structure and get an idea of how it looked in the past.
• They can be used in plastic surgery where the organs, face, limbs or any other portion of the body has been damaged and needs to be rebuilt.
• It can be used in airport security, where concealed shapes can be used for guessing whether a person is armed or is carrying explosives or not.
• It can also help in completing DNA sequences.

So, if you are planning to implement this technology, then you can rent the required infrastructure from E2E Networks and avoid investing in it. And if you plan to learn more about such topics, then keep a tab on the blog section of the website

https://tongtianta.site/paper/68922

https://github.com/natowi/3D-Reconstruction-with-Deep-Learning-Methods

October 18, 2022

### A Comprehensive Guide To Deep Q-Learning For Data Science Enthusiasts

For all data science enthusiasts who would love to dig deep, we have composed a write-up about Q-Learning specifically for you all. Deep Q-Learning and Reinforcement learning (RL) are extremely popular these days. These two data science methodologies use Python libraries like TensorFlow 2 and openAI’s Gym environment.

So, read on to know more.

What is Deep Q-Learning?

Deep Q-Learning utilizes the principles of Q-learning, but instead of using the Q-table, it uses the neural network. The algorithm of deep Q-Learning uses the states as input and the optimal Q-value of every action possible as the output. The agent gathers and stores all the previous experiences in the memory of the trained tuple in the following order:

State> Next state> Action> Reward

The neural network training stability increases using a random batch of previous data by using the experience replay. Experience replay also means the previous experiences stocking, and the target network uses it for training and calculation of the Q-network and the predicted Q-Value. This neural network uses openAI Gym, which is provided by taxi-v3 environments.

Now, any understanding of Deep Q-Learning   is incomplete without talking about Reinforcement Learning.

What is Reinforcement Learning?

Reinforcement is a subsection of ML. This part of ML is related to the action in which an environmental agent participates in a reward-based system and uses Reinforcement Learning to maximize the rewards. Reinforcement Learning is a different technique from unsupervised learning or supervised learning because it does not require a supervised input/output pair. The number of corrections is also less, so it is a highly efficient technique.

Now, the understanding of reinforcement learning is incomplete without knowing about Markov Decision Process (MDP). MDP is involved with each state that has been presented in the results of the environment, derived from the state previously there. The information which composes both states is gathered and transferred to the decision process. The task of the chosen agent is to maximize the awards. The MDP optimizes the actions and helps construct the optimal policy.

For developing the MDP, you need to follow the Q-Learning Algorithm, which is an extremely important part of data science and machine learning.

What is Q-Learning Algorithm?

The process of Q-Learning is important for understanding the data from scratch. It involves defining the parameters, choosing the actions from the current state and also choosing the actions from the previous state and then developing a Q-table for maximizing the results or output rewards.

The 4 steps that are involved in Q-Learning:

1. Initializing parameters – The RL (reinforcement learning) model learns the set of actions that the agent requires in the state, environment and time.
2. Identifying current state – The model stores the prior records for optimal action definition for maximizing the results. For acting in the present state, the state needs to be identified and perform an action combination for it.
3. Choosing the optimal action set and gaining the relevant experience – A Q-table is generated from the data with a set of specific states and actions, and the weight of this data is calculated for updating the Q-Table to the following step.
4. Updating Q-table rewards and next state determination – After the relevant experience is gained and agents start getting environmental records. The reward amplitude helps to present the subsequent step.

In case the Q-table size is huge, then the generation of the model is a time-consuming process. This situation requires Deep Q-learning.

Hopefully, this write-up has provided an outline of Deep Q-Learning and its related concepts. If you wish to learn more about such topics, then keep a tab on the blog section of the E2E Networks website.

https://analyticsindiamag.com/comprehensive-guide-to-deep-q-learning-for-data-science-enthusiasts/

https://medium.com/@jereminuerofficial/a-comprehensive-guide-to-deep-q-learning-8aeed632f52f

October 13, 2022

### GAUDI: A Neural Architect for Immersive 3D Scene Generation

The evolution of artificial intelligence in the past decade has been staggering, and now the focus is shifting towards AI and ML systems to understand and generate 3D spaces. As a result, there has been extensive research on manipulating 3D generative models. In this regard, Apple’s AI and ML scientists have developed GAUDI, a method specifically for this job.

An introduction to GAUDI

The GAUDI 3D immersive technique founders named it after the famous architect Antoni Gaudi. This AI model takes the help of a camera pose decoder, which enables it to guess the possible camera angles of a scene. Hence, the decoder then makes it possible to predict the 3D canvas from almost every angle.

What does GAUDI do?

GAUDI can perform multiple functions –

• The extensions of these generative models have a tremendous effect on ML and computer vision. Pragmatically, such models are highly useful. They are applied in model-based reinforcement learning and planning world models, SLAM is s, or 3D content creation.
• Generative modelling for 3D objects has been used for generating scenes using graf, pigan, and gsn, which incorporate a GAN (Generative Adversarial Network). The generator codes radiance fields exclusively. Using the 3D space in the scene along with the camera pose generates the 3D image from that point. This point has a density scalar and RGB value for that specific point in 3D space. This can be done from a 2D camera view. It does this by imposing 3D datasets on those 2D shots. It isolates various objects and scenes and combines them to render a new scene altogether.
• GAUDI also removes GANs pathologies like mode collapse and improved GAN.
• GAUDI also uses this to train data on a canonical coordinate system. You can compare it by looking at the trajectory of the scenes.

How is GAUDI applied to the content?

The steps of application for GAUDI have been given below:

• Each trajectory is created, which consists of a sequence of posed images (These images are from a 3D scene) encoded into a latent representation. This representation which has a radiance field or what we refer to as the 3D scene and the camera path is created in a disentangled way. The results are interpreted as free parameters. The problem is optimized by and formulation of a reconstruction objective.
• This simple training process is then scaled to trajectories, thousands of them creating a large number of views. The model samples the radiance fields totally from the previous distribution that the model has learned.
• The scenes are thus synthesized by interpolation within the hidden space.
• The scaling of 3D scenes generates many scenes that contain thousands of images. During training, there is no issue related to canonical orientation or mode collapse.
• A novel de-noising optimization technique is used to find hidden representations that collaborate in modelling the camera poses and the radiance field to create multiple datasets with state-of-the-art performance in generating 3D scenes by building a setup that uses images and text.

To conclude, GAUDI has more capabilities and can also be used for sampling various images and video datasets. Furthermore, this will make a foray into AR (augmented reality) and VR (virtual reality). With GAUDI in hand, the sky is only the limit in the field of media creation. So, if you enjoy reading about the latest development in the field of AI and ML, then keep a tab on the blog section of the E2E Networks website.