Huawei ML FAQ
How do I get an HCIA certificate? — AISeries — Episode #01
How do I get a Huawei HCIA certificate?
Well, not yet…but consider studying this material with me while I am preparing for the test:)
I annotated my errors from the HCIA-AI V3.0 Mock Exam :/
Index:01 — Do Pytorch and TensorFlow support GPU Accelerations?
02 — What Is ModelArts?
03 — What is the difference between ANN, CNN, and RNN?
04 — Is Highway Network a Recurrent Neural Network?
05 - What is the Vanishing Gradient Problem (vgp)?
06 — What are the TensorFlow ranks and tensors?
07 — PyTorch use Lua Programming Language?
08 — What is HUAWEI HiAI Foundation?
09 — What is MindSpore?
10 — What are the service types of Huawei cloud EI?
11 — What are some products of Huawei?
12 — what are the most prominent techniques of data preprocessing in machine learning?
13 — What is GAN in machine learning? How does a GAN work? Is GAN an ML’s Nirvana?
14 — What is ReLU in machine learning? Why is ReLU so good? Who invented ReLU?
Let’s get it on!
01Q&A— Do Pytorch and TensorFlow support GPU Accelerations?
Yes, they do.These are the two most popular frameworks in Deep Learning Area: PyTorch and TensorFlow.Googles' TensorFlow is a very powerful and mature deep learning library with strong visualization capabilities and several options to use for high-level model development. It has production-ready deployment options and support for mobile platforms. Facebook's PyTorch, on the other hand, is still a young framework with stronger community movement and it's more Python friendly.CUDA is a parallel computing platform and programming model developed by Nvidia for general computing on its own GPUs (graphics processing units). NVIDIA® CUDA enables developers to speed up compute-intensive applications by harnessing the power of GPUs for the parallelizable part of the computationThe NVIDIA® CUDA® Deep Neural Network library™ (cuDNN) is a GPU-accelerated library of primitives for deep neural networks.
cuDNN is a deep learning GPU acceleration library based on CUDA.The combination of CUDA and Nvidia GPUs dominates several application areas, including deep learning, and is a foundation for some of the fastest computers in the world.My card is a Nvidia GeForce GTX 1660 Ti w/ Max-Q Design. Nowadays, the top is a NVIDIA TESLA V100 TENSOR CORE GPU PCIe (link).Many deep learning frameworks either don’t support it or support it only as an afterthought once their CUDA support has been released.In this case, yes, both PyTorch and TensorFlow support GPU acceleration.
02Q&A— What Is ModelArts?
HUAWEI CLOUD ModelArts is a one-stop development platform for AI developers. With data preprocessing, semi-automated data labeling, distributed training, automated model building, and on-demand model deployment on the device, edge, and cloud, ModelArts helps AI developers quickly build models and efficiently manage the AI development lifecycle.ModelArts is suitable for AI developers with varying levels of development experience. Service developers can use ExeML (customized code-free model) to quickly build AI applications without coding. Beginners can directly use built-in algorithms to build AI applications. AI engineers can use multiple development environments to quickly compile code for modeling and application development.
03Q&A — What is the difference between ANN, CNN, and RNN?
Artificial Neural Network (ANN) is a group of multiple perceptrons or neurons at each layer. ANN is also known as a Feed-Forward Neural network because inputs are processed only in the forward direction.
This type of neural networks are one of the simplest variants of neural networks. They pass information in one direction, through various input nodes, until it makes it to the output node. The network may or may not have hidden node layers, making their functioning more interpretable;Convolutional Neural Network (CNN) are one of the most popular models used today. This neural network computational model uses a variation of multilayer perceptrons and contains one or more convolutional layers that can be either entirely connected or pooled. These convolutional layers create feature maps that record a region of image which is ultimately broken into rectangles and sent out for nonlinear processingConvolutional neural network is better than a feed-forward network since CNN has features parameter sharing and dimensionality reduction. Because of parameter sharing in CNN, the number of parameters is reduced thus the computations also decreased. The main intuition is the learning from one part of the image is also useful in another part of the image. Because of the dimensionality reduction in CNN, the computational power needed is reduced.Basically, Convolutional neural network are your best friend today to automatically do Feature Engineering (Feature Learning or Feature Extractions) without wasting too much time creating your own code and with no need to expertise in the field of feature engineering.Recurrent neural networks (RNN) are more complex. They save the output of processing nodes and feed the result back into the model (they did not pass the information in one direction only). This is how the model is said to learn to predict the outcome of a layer. Each node in the RNN model acts as a memory cell, continuing the computation and implementation of operations. If the network’s prediction is incorrect, then the system self-learns and continues working towards the correct prediction during backpropagation.
04Q&A — Is Highway Network a Recurrent Neural Network?
No, it doesn't.In machine learning, a highway network is an approach to optimizing networks and increasing their depth. Highway networks use learned gating mechanisms to regulate information flow, inspired (but not identical) by Long Short-Term Memory (LSTM) recurrent neural networks;Highway networks have been used as part of text sequence labeling and speech recognition tasks;The advantage of a Highway Network over the common deep neural networks is that solves or prevents partially the Vanishing Gradient Problem (for vgp, see Q&A#05), thus leading to easier to optimize neural networks.
05Q&A — What is the Vanishing Gradient Problem (vgp)?
The Vanishing Gradient Problem is an issue that sometimes arises when training machine learning algorithms through gradient descent. This most often occurs in neural networks that have several neuronal layers such as in a deep learning system, but also occurs in recurrent neural networks. The key point is that the calculated partial derivatives (see this post:) used to compute the gradient as one goes deeper into the network. Since the gradients control how much the network learns during training, if the gradients are very small or zero, then little to no training can take place, leading to poor predictive performance.
06Q&A —What are the TensorFlow ranks and tensors?
The TensorFlow library lets users define operations and functions over tensors as computational graphs. Tensors are a generalizable mathematical notation for multidimensional arrays holding data values, where the dimensionality of a tensor is typically referred to as its rank.We've worked mostly, so far, with tensors of rank zero to two. For instance, a scalar, a single number such as an integer or float, is a tensor of rank 0. A vector is a tensor of rank 1, and a matrix is a tensor of rank 2. But, it doesn't stop here. The tensor notation can be generalized to higher dimensions as well.To make the concept of a tensor more intuitive, consider the following figure, which represents tensors of ranks 0 and 1 in the first row, and tensors of ranks 2 and 3 in the second row:
07Q&A —PyTorch use Lua Programming Language?
No, it doesn't.Please, do not confuse Torch and PyTorch.Torch:A Tensor library like numpy, unlike numpy it has strong GPU support.
Lua is a wrapper for Torch (Yes! you need to have a good understanding of Lua), and for that you will need LuaRocks package manager.PyTorch:No need for the LuaRocks package manager, no need to write code in Lua. And because we are using Python, we can develop Deep Learning models with utmost flexibility. We can also exploit major Python packages likes scipy, numpy, matplotlib and Cython with PyTorch's own autograd.
08Q&A — What is HUAWEI HiAI Foundation?
HUAWEI HiAI is an open AI capability platform for smart devices (Mobile Phones). It encompasses a smart and comprehensive ecosystem based on an innovative "chip-device-cloud" three-layer open architecture, enabling developers to quickly and easily use the powerful AI capabilities on Huawei devices, and deliver a next-level smart app user experience.HUAWEI HiAI Foundation serves as a basic platform providing chip-level capabilities, and a high-performing computing environment characterized by low power consumption and all-scenario versatility, with broad-based support for more than 300 operators and features across the industry.HUAWEI HiAI Foundation connects smart services and computing chips to provide lightweight model computing capabilities. Supporting interconnection with frameworks such as Huawei's MindSpore, Google's TensorFlow, Facebook's PyTorch, UC Berkeley's Caffe, Paddle, Linux Foudation's ONNX and AndroidNN.HUAWEI HiAI Foundation provides for heterogeneous computing of multiple computing units in the chip, and offers efficient tool chains such as model quantization, model conversion, performance optimization, and maintenance and test.
09Q&A —What is MindSpore?
MindSpore is the AI computing framework. Fully developed by Huawei from the ground up, it implements on-demand collaboration across the cloud-edge-device. It provides unified APIs and end-to-end AI capabilities for model development, execution, and deployment in all scenarios.
10Q&A —What are the service types of Huawei cloud EI?
ModelArts
Intelligent Twins
EI Experience Hall
EI Service FamilyEI means Enterprise Intelligence.ModelArts is commented in question and answer Q&A#02;Industrial Intelligent Twins uses big data and AI technologies to provide a full series of services covering design, production, logistics, sales, and service. It helps enterprises build the leading position in the industry;Enterprise Intelligence Experience Hall or EI Experience Center is a service portfolio provides intelligent enable devices with the convenience of service as platform (check it out)[Enterprise Intelligence] EI Service Family are solutions, such as, Big Data, Essential Platform, Conversational Bot, Natural Language Processing, Automatic Speech, and so on...
11Q&A — What are some products of Huawei?
Atlas 200 DK AI Developer Kit
The Atlas 200 DK (model: 3000) is a high-performance AI application developer board that integrates the Ascend 310 AI processor to facilitate quick development and verification. It has been widely used in scenarios such as developer solution verification, higher education, and scientific research.
Ascend 310
This highly efficient, flexible, and programmable AI processor delivers 22 TOPS@INT8 and 11 TFLOPS@FP16 with just 8 W of power consumption. Ascend 310 integrates Huawei’s own Da Vinci architecture with abundant computing units, extending AI chip applications. As services accelerate their transformation to full intelligence, AI system performance excels while deployment costs fall.
Huawei Ascend 910
Claimed To Be Most Powerful AI Processor And MindSpore, An AI Computing Framework Launched
12Q&A — what are the most prominent techniques of data preprocessing in machine learning?
Data Standardization;
Missing Value;
Data Dimensionality Reduction;
Data Merging;
(i)Feature Discretization;
(ii)Statistical & Correlation Analysis;
(iii)Dimensionally Reduction Methods for Feature Extraction
(iv)Image Preprocessing Methods (Reconstruction, Segmentation, Sharpening);
(v)Data Mining Algorithms;
(vi)Patters Reconition;and much, much more ;-) Data preprocessing includes functionalities for (i) feature discretization, (ii) correlation analysis and statistical analysis to select clinical features that appear to be significant (feature selection), (iii) dimensionality reduction methods for extracting significant features from genetic data (e.g., transcriptomic), (iv) image preprocessing methods (e.g., reconstruction, segmentation, sharpening), (v) data modeling that involves the application of data mining algorithms on preprocessed medical data (e.g., laboratory measures) for (vi)pattern recognition, etc. (source)
13Q&A — What is GAN in machine learning? How does a GAN work? Is GAN an ML’s Nirvana?
Generative adversarial networks (GANs) are an exciting recent innovation in machine learning. GANs are generative models: they create new data instances that resemble your training data. For example, GANs can create images that look like photographs of human faces, even though the faces don't belong to any real person.GANs consists of two networks, a Generator G(x), and a Discriminator D(x). They both play an adversarial game where the generator tries to fool the discriminator by generating data similar to those in the training set. The Discriminator tries not to be fooled by identifying fake data from real data.GANs have a number of common failure modes. All of these common problems are areas of active research. While none of these problems have been completely solved, I'll mention some things that I've annotated from HCIA-AI V3.0 Mock Exam:. An over high learning rate leads to model divergence;
. If the discriminator is too well trainned the generator cannot be trainned;
. The model colapse results in insufficient diversified imagem;
. If data set is small, overfitting is the inssue :/
14Q&A — What is ReLU in machine learning? Why is ReLU so good? Who invented ReLU?
ReLU is the most commonly used activation function in neural networks, especially in CNNs (see Q&A#03). If you are unsure what activation function to use in your network, ReLU is usually a good first choice.ReLU and its variants come standard with most frameworks: TensorFlow, Keras, PyTorch, etc.The biggest advantage of ReLu is indeed non-saturation of its gradient (see vgp Q&A#05), which greatly accelerates the convergence of stochastic gradient descent compared to the sigmoid / tanh functions (paper by Krizhevsky et al);The Rectified Linear Unit (ReLU) activation function was proposed by Nair and Hinton 2010, and ever since, has been the most widely used activation function for deep learning applications with state-of-the-art results to date;In summary:
. ReLu is used to mitigating the problem of Vanishing Gradient Problem - vgp (see Q&A#05 above);
. It saves calculation as well :)
That’s all!
In the next episode more tips on Huawei Technologies.
I hope these study materials can help you (and myself:) get the Huawei certificate HCIA easily.
We might get a good grade!
I will fight for it. I will not give up. I will reach my goal. Absolutely nothing will stop me.
Be my guest!
Credits & References
HCIA-AI V3.0 Mock Exam by https://ilearningx.huawei.com/portal/dashboard/index
Related Posts
00#Episode — AISeries — ML — Machine Learning Intro — What Is It and How It Evolves Over Time?
01#Episode — AISeries — Huawei ML FAQ — How do I get an HCIA certificate? (this one:)
02#Episode — AISeries — Huawei ML FAQ Again — More annotation from Huawei Mock Exam
03#Episode — AISeries — AI In Graphics — Getting Intuition About Complex Math & More
04#Episode — AISeries — Huawei ML FAQ — Advanced — Even More annotation from Huawei Mock Exam
05#Episode — AISeries — SVM — Credit Card — Start to Finished — A Complete Colab Notebook Using the Default of Credit Card Clients Data Set from UCI
06#Episode — AISeries — SVM — Breast Cancer — Start to Finished — A Complete Colab Notebook Using the Default of Credit Card Clients Data Set from UCI
07#Episode — AISeries — SVM — Cupcakes or Muffins? — Start To Finished — Based on Alice Zhao post
Creativity is a skill that’s in demand. Boost yours with these tips:)