The field of DL has matured a lot in the last decade and changed a lot in the last few years. New architectures scaled to be larger/deeper, take advantage of a large number of datasets and parallel computing power.
Supervised DL methods, namely, CNNs and RNNs, are the natural choice for researchers in the automotive domain.
Important aspects such as compute power requirements, model transparency, and interpretability, model compliance with vehicle safety standards, all of which are expected to appreciably impact the adoption rate of DL in the automotive industry.
With the development of autonomous development
technology, the need for additional applications to be used
inside and outside the vehicle is increasing. As a result of the
literature review, many applications have been developed to
display vehicle data directly on the monitor, with reflections
on glass, and on hardware devices. These applications have
been developed only for a defined problem and for a
particular autonomous system. In this study, a basic
autonomous vehicle software infrastructure and mobile
Augmented Reality application that can work on Android
devices have been developed. The Mobile Augmented Reality
app serves inside and outside the vehicle. In addition, this
application has been shown to support multiple autonomous
system infrastructures.
SOFTWARE AND HARDWARE DESIGN CHALLENGES IN AUTOMOTIVE EMBEDDED SYSTEMVLSICS Design
Modern automotives integrate large amount of electronic devices to improve the driving safety and comfort. This growing number of Electronic Control Units (ECUs) with sophisticated software escalates the vehicle system design complexity. In this paper we explain the complexity of ECUs in terms of hardware and software and also we explore the possibility of Common Object Request Broker Architecture (CORBA) architecture for the integration of add-on software in ECUs. This reduces the complexity of the embedded system in vehicles and eases the ECU integration by reducing the total number of ECUs in the vehicles.
ITCamp 2019 - Silviu Niculita - Supercharge your AI efforts with the use of A...ITCamp
Microsoft "Automated Machine Learning" (AutoML) is an amazing toolkit now available on Azure that's really starting to ramp up.
In a nutshell, it is an automated service that identifies the best machine learning pipelines for labeled data ... it dramatically frees up time for experienced practitioners and gives a tremendous boost to in productivity engineers at the start of their ML journey.
With the development of autonomous development
technology, the need for additional applications to be used
inside and outside the vehicle is increasing. As a result of the
literature review, many applications have been developed to
display vehicle data directly on the monitor, with reflections
on glass, and on hardware devices. These applications have
been developed only for a defined problem and for a
particular autonomous system. In this study, a basic
autonomous vehicle software infrastructure and mobile
Augmented Reality application that can work on Android
devices have been developed. The Mobile Augmented Reality
app serves inside and outside the vehicle. In addition, this
application has been shown to support multiple autonomous
system infrastructures.
SOFTWARE AND HARDWARE DESIGN CHALLENGES IN AUTOMOTIVE EMBEDDED SYSTEMVLSICS Design
Modern automotives integrate large amount of electronic devices to improve the driving safety and comfort. This growing number of Electronic Control Units (ECUs) with sophisticated software escalates the vehicle system design complexity. In this paper we explain the complexity of ECUs in terms of hardware and software and also we explore the possibility of Common Object Request Broker Architecture (CORBA) architecture for the integration of add-on software in ECUs. This reduces the complexity of the embedded system in vehicles and eases the ECU integration by reducing the total number of ECUs in the vehicles.
ITCamp 2019 - Silviu Niculita - Supercharge your AI efforts with the use of A...ITCamp
Microsoft "Automated Machine Learning" (AutoML) is an amazing toolkit now available on Azure that's really starting to ramp up.
In a nutshell, it is an automated service that identifies the best machine learning pipelines for labeled data ... it dramatically frees up time for experienced practitioners and gives a tremendous boost to in productivity engineers at the start of their ML journey.
Autonomous vehicles: A study of implementation and security IJECEIAES
Autonomous vehicles have been invented to increase the safety of transportation users. These vehicles can sense their environment and make decisions without any external aid to produce an optimal route to reach a destination. Even though the idea sounds futuristic and if implemented successfully, many current issues related to transportation will be solved, care needs to be taken before implementing the solution. This paper will look at the pros and cons of implementation of autonomous vehicles. The vehicles depend highly on the sensors present on the vehicles and any tampering or manipulation of the data generated and transmitted by these can have disastrous consequences, as human lives are at stake here. Various attacks against the different type of sensors on-board an autonomous vehicle are covered.
A Model for Car Plate Recognition & Speed Tracking (CPR-STS) using Machine Le...CSCJournals
The transportation challenges experienced in major cities as a result of influx of people in search of greener pastures is increasing on a daily basis. This results in an increase in the number of cars plying and competing for driving space on narrow roads. Many drivers violate traffic laws as a result of this and how to prosecute them without chasing them remains an issue to be addressed. Therefore, this research presents a model that can be used to solve this challenge using machine learning algorithms. The model consists of recognition modules such as image acquisition, Gaussian blur, localization of car plate, character segmentation and optical character recognition of car plate. K-NN Algorithm was used for training licensed plate font type spanning A-Z and 0-9 while the speed tracking module used a camera which is automatically self-initiated to track the speed of any moving object within its range of focus. The performance of the model was evaluated using metrics such as recognition accuracy, positive prediction value, negative prediction value, specificity and sensitivity. A tracking accuracy of 82% was achieved.
Mobile devices such as smart phones, have gained enormous popularity over the last few years. Smart phones are now capable of supporting a large number of applications, many of which demand an ever increasing computational power. However, Battery life of mobile devices remains a key limiting factor in the design of mobile applications. This paper firstly emphasizes on reviewing various techniques used to enhance battery life of smart phones. At the end, on the basis of comparative studies, suggestions are given to extend the energy efficiency of smart phones.
With the spotlight shining brighter on autonomous driving, 5G and cellular-vehicle-to-everything (CV2X) breakthroughs grab the headlines when it comes to reporting how to better test self-driving cars. Behind this exciting and somewhat glamourous buzz is a robust automotive electronics manufacturing services (EMS) industry that’s physicalizing the technology innovations.
Automotive Cybersecurity Challenges for Automated Vehicles: Jonathan PetitSecurity Innovation
July 2016: Jonathan Petit, Principal Scientist at Security Innovation, discusses cybersecurity challenges for automated vehicles at the Automotive Vehicles Symposium.
Vehicle accidents are by all accounts appalling and frightening occasions occurring which cause various deaths. As the number of accidents per year is increasing tremendously and so the lives affected by accidents. There are traditional ways to help the needy or the victim that is informing the right authority but needs assistance or help from others, but this tends to take ample of time and due to it could cost lives. So there is a need to develop an accident detection system that would detect and alert the proper authorities about the accident. The sudden assistance to the alert would in return lead to saving as many lives as possible. Many researchers have analyzed this technique using Convolutional neural network, HDNN, RCNN, etc. This paper will give us an overview of various techniques or methods that are used to detect accidents.
Multiagent multiobjective interaction game system for service provisoning veh...redpel dot com
Multiagent multiobjective interaction game system for service provisoning vehicular cloud
for more ieee paper / full abstract / implementation , just visit www.redpel.com
A fully automated, self-driving car can perceive its environment, determine the optimal route, and drive
unaided by human intervention for the entire journey. Connected autonomous vehicles (CAVs) have the
potential to drastically reduce accidents, travel time, and the environmental impact of road travel. Such
technology includes the use of several sensors, various algorithms, interconnected network connections,
and multiple auxiliary systems. CAVs have been subjected to attacks by malicious users to gain/deny
control of one or more of its various systems. Data security and data privacy is one such area of CAVs that
has been targeted via different types of attacks. The scope of this study is to present a good background
knowledge of issues pertaining to different attacks in the context of data security and privacy, as well
present a detailed review and analysis of eight very recent studies on the broad topic of security and
privacy related attacks. Methodologies including Blockchain, Named Data Networking, Intrusion
Detection System, Cognitive Engine, Adversarial Objects, and others have been investigated in the
literature and problem- and context-specific models have been proposed by their respective authors.
This presentation talks about Software Defined Vehicles, Automotive Standards including Cyber Security and Safety, Agile Methods like SAFe/Less , Continuous Delivery best practices.
The choice of technology within the autonomous vehicle ecosystem drives the business model. The article presents and highlights a few choices and their implications
Autonomous vehicles: A study of implementation and security IJECEIAES
Autonomous vehicles have been invented to increase the safety of transportation users. These vehicles can sense their environment and make decisions without any external aid to produce an optimal route to reach a destination. Even though the idea sounds futuristic and if implemented successfully, many current issues related to transportation will be solved, care needs to be taken before implementing the solution. This paper will look at the pros and cons of implementation of autonomous vehicles. The vehicles depend highly on the sensors present on the vehicles and any tampering or manipulation of the data generated and transmitted by these can have disastrous consequences, as human lives are at stake here. Various attacks against the different type of sensors on-board an autonomous vehicle are covered.
A Model for Car Plate Recognition & Speed Tracking (CPR-STS) using Machine Le...CSCJournals
The transportation challenges experienced in major cities as a result of influx of people in search of greener pastures is increasing on a daily basis. This results in an increase in the number of cars plying and competing for driving space on narrow roads. Many drivers violate traffic laws as a result of this and how to prosecute them without chasing them remains an issue to be addressed. Therefore, this research presents a model that can be used to solve this challenge using machine learning algorithms. The model consists of recognition modules such as image acquisition, Gaussian blur, localization of car plate, character segmentation and optical character recognition of car plate. K-NN Algorithm was used for training licensed plate font type spanning A-Z and 0-9 while the speed tracking module used a camera which is automatically self-initiated to track the speed of any moving object within its range of focus. The performance of the model was evaluated using metrics such as recognition accuracy, positive prediction value, negative prediction value, specificity and sensitivity. A tracking accuracy of 82% was achieved.
Mobile devices such as smart phones, have gained enormous popularity over the last few years. Smart phones are now capable of supporting a large number of applications, many of which demand an ever increasing computational power. However, Battery life of mobile devices remains a key limiting factor in the design of mobile applications. This paper firstly emphasizes on reviewing various techniques used to enhance battery life of smart phones. At the end, on the basis of comparative studies, suggestions are given to extend the energy efficiency of smart phones.
With the spotlight shining brighter on autonomous driving, 5G and cellular-vehicle-to-everything (CV2X) breakthroughs grab the headlines when it comes to reporting how to better test self-driving cars. Behind this exciting and somewhat glamourous buzz is a robust automotive electronics manufacturing services (EMS) industry that’s physicalizing the technology innovations.
Automotive Cybersecurity Challenges for Automated Vehicles: Jonathan PetitSecurity Innovation
July 2016: Jonathan Petit, Principal Scientist at Security Innovation, discusses cybersecurity challenges for automated vehicles at the Automotive Vehicles Symposium.
Vehicle accidents are by all accounts appalling and frightening occasions occurring which cause various deaths. As the number of accidents per year is increasing tremendously and so the lives affected by accidents. There are traditional ways to help the needy or the victim that is informing the right authority but needs assistance or help from others, but this tends to take ample of time and due to it could cost lives. So there is a need to develop an accident detection system that would detect and alert the proper authorities about the accident. The sudden assistance to the alert would in return lead to saving as many lives as possible. Many researchers have analyzed this technique using Convolutional neural network, HDNN, RCNN, etc. This paper will give us an overview of various techniques or methods that are used to detect accidents.
Multiagent multiobjective interaction game system for service provisoning veh...redpel dot com
Multiagent multiobjective interaction game system for service provisoning vehicular cloud
for more ieee paper / full abstract / implementation , just visit www.redpel.com
A fully automated, self-driving car can perceive its environment, determine the optimal route, and drive
unaided by human intervention for the entire journey. Connected autonomous vehicles (CAVs) have the
potential to drastically reduce accidents, travel time, and the environmental impact of road travel. Such
technology includes the use of several sensors, various algorithms, interconnected network connections,
and multiple auxiliary systems. CAVs have been subjected to attacks by malicious users to gain/deny
control of one or more of its various systems. Data security and data privacy is one such area of CAVs that
has been targeted via different types of attacks. The scope of this study is to present a good background
knowledge of issues pertaining to different attacks in the context of data security and privacy, as well
present a detailed review and analysis of eight very recent studies on the broad topic of security and
privacy related attacks. Methodologies including Blockchain, Named Data Networking, Intrusion
Detection System, Cognitive Engine, Adversarial Objects, and others have been investigated in the
literature and problem- and context-specific models have been proposed by their respective authors.
This presentation talks about Software Defined Vehicles, Automotive Standards including Cyber Security and Safety, Agile Methods like SAFe/Less , Continuous Delivery best practices.
The choice of technology within the autonomous vehicle ecosystem drives the business model. The article presents and highlights a few choices and their implications
The Autonomous car is about to enter the mass-market. The question is not about when it will happen but in which conditions, under which form or who will be the first car manufacturer to release an efficient and reliable final product. Entirely unexpected ways to deal with building up the AI frameworks for self-driving vehicles exist and most of them are horribly best in class and with extremely high equipment needs. The appropriate response presented during this paper proposes the AI fundamentally based framework to be as simple as conceivable with low equipment needs. A straight forward three layers profound, totally associated neural system was prepared to outline pictures from a forward looking QVGA camera to directing orders. Upheld an information picture the neural system should settle on one among the four offered orders (Forward, Left, Right or Stop). With least of the instructing information (250 pictures) the framework figured out how to follow the street ahead and keep in its path.The framework precisely learns essential street alternatives with exclusively the directing point in light of the fact that the contribution from the human driver. it had been near explicitly prepared to watch lines out and about. Contrasted with rather progressively confounded arrangements like express decay of the issue, similar to path identification and the board and convolutional neural systems simply like the conclusion to complete the process of learning arranged by the N-Vidia this technique demonstrated to be amazingly solid and affordable. we will in general attempt to demonstrate that this methodology would bring about better and lower equipment necessities so making the occasion of oneself driving vehicles simpler and more financially savvy. Simple counterfeit neural system, much the same as the one gave during this paper, is sufficient for relatively muddled technique like path keeping.
Using the Open Source VS Code Editor with the HPCC Systems PlatformHPCC Systems
As part of the 2018 HPCC Systems Summit Community Day event:
Up first, Matt Butler, Kennesaw State University, briefly discusses his poster, The Future of Automotive Telemetry: Assessing Inherent Risk Implications and Cyber Security Vulnerabilities.
Following, Arjuna Chala presents his breakout session in the User Interfaces track.
Are you a fan of the VS Code Editor for building and debugging your code? Join me as I demonstrate how this can be used as a modern, extensible alternative to the ECL IDE and why most developers choose the VS Code editor due to its benefits of being cross-platform, multi-language support, and open source.
Arjuna Chala is Sr. Director of Special Projects for the HPCC Systems® platform at LexisNexis Risk Solutions®. With almost 20 years of experience in software design, Arjuna leads the development of next generation big data capabilities including creating tools around exploratory data analysis, data streaming and business intelligence. Arjuna strives to understanding new technologies and bring innovative applications and design to the HPCC System platform. Dedicated to development excellence, Arjuna served as a key member of the team to bring the HPCC Systems platform to the open source community. In his work with HPCC Systems community leaders and system integrator partners, Arjuna’s efforts have contributed to the spread of HPCC Systems technology into the enterprise domestically as well as the international markets of China, Brazil, Europe and India. Arjuna has a BS in Computer Science from RVCE, Bangalore University.
Smart City solution providers will face challenges in increasing network load due to the huge amounts of video data flowing through their networks. For cost-effective analytics, distributed architecture with user control is just the right solution required. In Smart Cities with varying applications of video analytics solutions in fields such as security systems, utilities operators, and emergency response systems, it gives users a simple way to pick the feed they would like, instrument the analysis they want, and report the way they require in a simple-configurable manner.
ATMOSPHERE was invited to be a speaker at Think Milano event, on 6th June from 14.30 to 17.30, to join a panel discussion called “L’infrastruttura cloud ready protagonista del future” on how cloud infrastructures are important for different market sectors.
Similar to Deep learning in automation industry (20)
06-04-2024 - NYC Tech Week - Discussion on Vector Databases, Unstructured Data and AI
Round table discussion of vector databases, unstructured data, ai, big data, real-time, robots and Milvus.
A lively discussion with NJ Gen AI Meetup Lead, Prasad and Procure.FYI's Co-Found
Quantitative Data AnalysisReliability Analysis (Cronbach Alpha) Common Method...2023240532
Quantitative data Analysis
Overview
Reliability Analysis (Cronbach Alpha)
Common Method Bias (Harman Single Factor Test)
Frequency Analysis (Demographic)
Descriptive Analysis
Explore our comprehensive data analysis project presentation on predicting product ad campaign performance. Learn how data-driven insights can optimize your marketing strategies and enhance campaign effectiveness. Perfect for professionals and students looking to understand the power of data analysis in advertising. for more details visit: https://bostoninstituteofanalytics.org/data-science-and-artificial-intelligence/
06-04-2024 - NYC Tech Week - Discussion on Vector Databases, Unstructured Data and AI
Discussion on Vector Databases, Unstructured Data and AI
https://www.meetup.com/unstructured-data-meetup-new-york/
This meetup is for people working in unstructured data. Speakers will come present about related topics such as vector databases, LLMs, and managing data at scale. The intended audience of this group includes roles like machine learning engineers, data scientists, data engineers, software engineers, and PMs.This meetup was formerly Milvus Meetup, and is sponsored by Zilliz maintainers of Milvus.
2. TABLE OF CONTENTS
ABSTRACT
INTRODUCTION
LITERATURE SURVEY:-
1. ARCHITECTURE OF DEEP LEARNING MODELS
2. DEEP LEARNING BASED AUTOMOTIVE USE CASES
CONCLUSION
FUTURE SCOPE
REFERENCES
3. ABSTRACT
One of the most exciting technology breakthroughs in the last few years has been the rise of deep learning.
The increasing scale of data, computational power and the associated algorithmic innovations are the main
drivers for the progress we see in this field.
These developments also have a huge potential for the automotive industry and therefore the interest in deep
learning-based technology is growing.
A lot of the product innovations, such as self-driving cars, parking and lane-change assist or safety functions,
such as autonomous emergency braking, are powered by deep learning algorithms.
Deep learning is poised to offer gains in performance and functionality for most ADAS (Advanced Driver
Assistance System) solutions, vehicle dynamics application, vehicle inspection/heath monitoring, automated
driving and data-driven product development.
4. INTRODUCTION
From voice assistants to self-driving cars, deep learning (DL) is redefining the way we interact with
machines.
DL has been able to achieve breakthroughs in historically difficult areas of machine learning such as text-
to-speech conversion, image classification, and speech recognition.
Yann LeCun developed convolutional neural network (CNN) in 1980’s. It could classify handwritten digits
[2] with good speed and accuracy and were widely deployed to read over 10% of all the cheques in the
United States.
DL has matured a lot in the last decade and changed a lot in the last few years.
The “deep” in deep learning stands for the many layers in the neural network that contribute to a model
of the data.
5. There are four key factors that are driving the progress and uptake of DL:
More Compute Power: e.g. graphics processing unit (GPUs), tensor processing unit (TPUs)
Organized Large Datasets: e.g. ImageNet
Better Algorithms: e.g. activation functions like ReLU , optimization schemes
Software & Infrastructure: e.g. Git, Robot Operating System (ROS), Tensor Flow, Keras
1
2
3
4
6. LITERATURE SURVEY
• Automotive industry has been a generator of large volumes of data.
• Data are collected from suppliers during the design,
development and testing of various subsystems, real-
time data are generated by hundreds of on-vehicle
sensors, system faults are recorded by the on-board
diagnostics (OBD), and information is collected about
the vehicle servicing history and customer preferences
at dealerships/service centers.
• Single autonomous vehicle will generate 4,000 GB
of data each day
7. ARCHITECTURE OF DEEP LEARNING MODELS
Conventional machine learning models are trained on features extracted from the raw data using
feature extraction and feature engineering techniques.
Designing a feature extractor that can successfully transform raw data into a suitable feature vector
requires considerable domain expertise.
DL solves the representation problem for unstructured data.
The key differentiating point of DL is that the model learns useful representations and features
automatically, directly from the raw data, bypassing this manual and difficult step of hand engineering
the feature vector.
8. Most commonly used libraries for training DL models include:
TensorFlow, a framework originating from Google Research
Keras, a DL library originally built by Francois Chollet and recently incorporated in TensorFlow
Caffe, a DL framework, originally developed at University of California, Berkeley
Theano, a Python library and optimizing compiler primarily developed by the Montreal Institute for
Learning Algorithms (MILA) at the Université de Montréal
Microsoft Cognitive Toolkit, previously known as CNTK
Pytorch, a framework associated with Facebook Research
1.
3.
6.
2.
4.
5.
10. • Training a DL algorithm is a computationally intensive process. It involves repetitive
arithmetic operations, requiring multi-core computing power due to the limited number of
arithmetic logic units (ALUs) on CPUs.
• Graphical processing units (GPUs) are designed for repetitive arithmetic operations. They
have thousands of ALUs and are naturally more suited for DL applications.
• It takes the
power of 16
CPUs to match
the power of 1
GPU
11. For the GPUs, NVIDIA introduced CUDA and cuDNN, which allow the use of popular languages such as C, C++,
Fortran, Python and MATLAB and express parallelism through extensions in the form of a few basic keywords.
For CPUs, Intel introduced the Math Kernel Library (MKL) which is optimized to speed up arithmetic operations.
A number of companies such as NVIDIA,
Mobileye, Intel,
Qualcomm, Apple and Samsung (Table II) are
developing specialized chips to enable
embedded implementation of real time DL
applications in smartphones, cameras, robots
and self-driving cars.
12. DEEP LEARNING BASED AUTOMOTIVE USE CASES
The selected use-cases are grouped under four key topics:
1.
2.
3.
4.
Virtual sensing for vehicle dynamics
Vehicle inspection/health monitoring
Automated driving
Data-driven product development
13. Virtual sensing for
vehicle dynamics
It focuses on the so-called soft/virtual sensors
that can be used for estimating relevant vehicle
dynamic and tire-related states.
Effective implementation of automotive stability control
systems largely depends on accurate vehicle dynamic state
information.
Knowledge of certain tire-vehicle dynamic states is of immense
importance for vehicle motion and control. Vehicle sideslip
angle (β) and the tire-road friction coefficient (μ) are probably
the two most prominent vehicle states of interest.
Vehicle sideslip angle is a strong indicator of the
vehicle’s
stability.
14. Vehicle inspection/health
monitoring
Vehicle repair workshops and insurance companies are beginning to tap into
the potential of DL.
Vision based algorithms are also being proposed as an effective approach in
preventing fraud claims as well as meeting the requirement to speed up the
insurance claim pre-processing.
DL technology is being used to automate visual tasks, such as inspecting the
damage to a vehicle. This approach provides drivers and insurance
companies with crucial information about the extent of the damage and also
presents tremendous opportunities to reduce the cost of manual labor.
Advanced predictive maintenance enabled by DL techniques is one of the most widely accepted benefits of the fourth
industrial revolution.
With predictive maintenance, faults can be reliably detected before they occur based on
data anomalies.
15. AUTOMATED DRIVING
Most established OEMs as well as upcoming startups and
mobility service providers like Uber, Cruise and Waymo
are working on building autonomous vehicles (AVs).
These AVs use sensors like camera, radar, LiDAR,
generating massive amounts of data every second.
16. Unlike traditional vehicles that mostly employed a model based approach with hand-coded decision logic for
vehicle control, most AVs typically leverage an end-to-end DL framework for vehicle control.
In an end-to-end framework, the entire decision-making logic is encoded into a neural network space that
produces the driving decisions.
Apart from vehicle control, end-to-end learning has also been used for other functions such driver distraction
recognition.
17. DATA DRIVEN PRODUCT
DEVELOPMENT
AI is expected to play a key role in data-driven product development.
Understanding the potential threat and developing effective ways to prevent, resolve, and
eliminate recalls is going to be critical.
Carmakers are investing heavily in technology that will help them track the performance of their
products from cradle to grave, and thus better anticipate and mitigate the risk of massive recalls.
18. CONCLUSION
The field of DL has matured a lot in the last decade, and changed a lot in the last few years. New architectures
scaled to be larger/deeper, take advantage of large number of datasets and parallel computing power.
Supervised DL methods, namely, CNNs and RNNs, are the natural choice for researchers in the automotive
domain.
As an outcome of this survey, novel uses cases have been identified and bucketed into three categories:
(1) virtual sensing for vehicle dynamics
(2) vehicle inspection/health monitoring and
(3) automated driving
Important aspects such as compute power requirements, model transparency and interpretability, model
compliance with vehicle safety standards, all of which are expected to appreciably impact the adoption rate of
DL in the automotive industry.
19. FUTURE SCOPE
In the future, we will:
Investigate distributed deep learning systems to improve training times for more
complex networks and larger data sets.
Assess and curate available datasets for computer vision use cases in the domain of
autonomous driving.
Evaluate natural understanding deep learning models (e. g., sequence-to-sequence
learning).
20. REFERENCES
[1] Y. Lecun, L. Bottou, Y. Bengio, and P. Haffner, "Gradient-based learning applied to document
recognition," Proc. IEEE, vol. 86, no. 11, pp. 2278–2324, Nov. 1998, doi: 10.1109/5.726791.
[2] Y. LeCun et al., "Backpropagation applied to handwritten zip code recognition," Neural
Computation, vol. 1, no. 4, pp. 541–551, Dec. 1989, doi: 10.1162/neco.1989.1.4.541.
[3] A. Cetinsoy, F. J. Martin, J. A. Ortega, and P. Petersen, "The past, present, and future of machine
learning APIs," in JMLR: Workshop Conf. Proc., 2016, pp. 43–49.
[4] F. Chollet and J. J. Allaire, Deep Learning with R. Greenwich, CT, USA: Manning Publications
Company, 2018.
[5] K. Sato, C. Young, and D. Patterson, "An in-depth look at Google’s first Tensor Processing Unit
(TPU)," Google Cloud Big Data and Machine Learning Blog, 2017.