Machine Learning Project Lifecycle
2 likes443 views
The document outlines the machine learning project lifecycle, focusing on deep neural networks and their training, as discussed in a presentation at Google I/O 2018. It covers essential stages including data collection, preparation, analysis, training, evaluation, and deployment, along with practical applications in various sectors like healthcare and financial trading. Additionally, it provides insights on resources for learning and applying machine learning techniques, including programming languages, frameworks, and relevant APIs.
1 of 57
Downloaded 29 times
![4. Train
30
ANN
𝑎 *
[,]
𝑎 .
[,]
𝑎 ,
[,]
𝑎,
[*]
𝑦(
𝑧,
,
= 𝑤,
, 2
𝑥 + 𝑏,
[,]
, 𝑎 ,
[,]
= 𝑔(𝑧,
,
)
𝑧*
,
= 𝑤*
, 2
𝑥 + 𝑏*
[,]
, 𝑎 *
[,]
= 𝑔(𝑧*
,
)
𝑧.
,
= 𝑤.
, 2
𝑥 + 𝑏.
[,]
, 𝑎 .
[,]
= 𝑔(𝑧.
,
)
𝑧,
*
= 𝑤,
* 2
𝑥 + 𝑏,
[*]
, 𝑎 ,
[*]
= 𝑔(𝑧,
*
)
Regularization
term
x1
x2
+
𝜆
2𝑚
:: 𝑊<
=*
>
=?,
@
<?,
min
D,E
𝐽(𝑊, 𝑏)](https://image.slidesharecdn.com/googleioextendedmlprojectlifecycle-last-180513231818/85/Machine-Learning-Project-Lifecycle-30-320.jpg)
![43
Output
Layer
Input
Layer
Hidden
Layer
Activation Functions
use linear when y is real (regression)
use ReLU when y is real positive (regression)
use sigmoid when y in {0,1} (2 classes)
use softmax when y in {0,1,2, ...} (K classes)
use tanh, ReLU or leaky ReLU
𝑎 *
[,]
𝑎 .
[,]
𝑎 ,
[,]
𝑎,
[*]
𝑦(](https://image.slidesharecdn.com/googleioextendedmlprojectlifecycle-last-180513231818/85/Machine-Learning-Project-Lifecycle-43-320.jpg)
Ad
Recommended
Machine learning
Machine learningADARSHMISHRA126 This document provides an overview of machine learning. It defines machine learning as a form of artificial intelligence that allows systems to automatically learn and improve from experience without being explicitly programmed. The document then discusses why machine learning is important, how it works by exploring data and identifying patterns with minimal human intervention, and provides examples of machine learning applications like autonomous vehicles. It also summarizes the main types of machine learning: supervised learning, unsupervised learning, reinforcement learning, and deep learning. Finally, it distinguishes machine learning from deep learning and defines data science.
Introduction to Machine Learning
Introduction to Machine LearningKmPooja4 This document summarizes Pooja's seminar presentation on machine learning. It introduces machine learning and compares it to traditional programming. It describes the main types of machine learning: supervised learning which uses labeled data to make predictions, unsupervised learning which finds patterns in unlabeled data, and reinforcement learning where an agent learns from feedback. The document discusses concepts like classification, regression, and feedback in machine learning systems. It also outlines some applications and concludes that machine learning can improve lives by advancing technology.
Big Data Analytics
Big Data AnalyticsGhulam Imaduddin The document discusses big data analytics, focusing on its impact on customer engagement in the telecommunications industry. It outlines challenges, technologies, methodologies, and tools used in big data analysis while highlighting examples and case studies in sentiment analysis and descriptive analytics using platforms like Hadoop and Spark. The document emphasizes the importance of analytical skills, data sources, and the role of a data scientist in leveraging big data effectively.
Machine learning
Machine learning Saurabh Agrawal The document outlines an Azure Machine Learning meetup focusing on the fundamentals of machine learning, including types of algorithms and tools. It explores supervised and unsupervised learning, provides practical demonstrations, and discusses the Azure ML platform's capabilities for predictive analytics. Key topics include classification, clustering, data preparation, and the complete data science process from problem definition to model deployment.
Intro/Overview on Machine Learning Presentation
Intro/Overview on Machine Learning PresentationAnkit Gupta This document provides an overview of a presentation on machine learning given at Gurukul Kangri University in 2017. It defines machine learning as a field that allows computers to learn without being explicitly programmed. It discusses different machine learning algorithms including supervised learning, unsupervised learning, and semi-supervised learning. Examples of applications of machine learning discussed include data mining, natural language processing, image recognition, and expert systems. The document also contrasts artificial intelligence, machine learning, and deep learning.
BIG DATA AND MACHINE LEARNING
BIG DATA AND MACHINE LEARNINGUmair Shafique The document discusses big data and machine learning, defining big data as vast and complex datasets that challenge traditional management tools, while highlighting its advantages and challenges. It also defines machine learning as an AI application that learns from data, presenting real-world applications and various types, such as supervised and unsupervised learning. The document emphasizes the relationship between big data and machine learning, outlining a framework for processing big data using machine learning techniques and presenting research challenges in the field.
introduction to data mining tutorial
introduction to data mining tutorial Salah Amean The document provides an overview of data mining and its significance, outlining its evolution, functionalities, and methodologies across various disciplines. It details the course structures related to data mining, highlighting major algorithms, popular issues in the field, and the transformative impact of data mining on business intelligence and decision-making. The introduction emphasizes the necessity of data mining in an era of exponential data growth and the challenges it presents in knowledge extraction from diverse data types.
Machine Learning and Real-World Applications
Machine Learning and Real-World ApplicationsMachinePulse The document discusses machine learning and its importance in handling the vast amounts of data generated daily, emphasizing its role in knowledge discovery. It outlines various learning methods, including supervised and unsupervised learning, and details algorithms for classification, regression, and clustering, such as k-nearest neighbors and neural networks. Additionally, it highlights real-world applications of machine learning, including recommender systems, medical diagnosis, and fraud detection.
Deep Learning Explained
Deep Learning ExplainedMelanie Swan This document summarizes Melanie Swan's presentation on deep learning. It began with defining key deep learning concepts and techniques, including neural networks, supervised vs. unsupervised learning, and convolutional neural networks. It then explained how deep learning works by using multiple processing layers to extract higher-level features from data and make predictions. Deep learning has various applications like image recognition and speech recognition. The presentation concluded by discussing how deep learning is inspired by concepts from physics and statistical mechanics.
ML Basics
ML BasicsSrujanaMerugu1 The document provides a comprehensive overview of machine learning (ML), detailing its applications, methodologies, and the development lifecycle. It emphasizes the importance of correctly framing business problems for ML solutions, selecting appropriate target variables, and understanding data integration and evaluation metrics. Additionally, the document discusses the benefits and challenges of using ML in real-world scenarios, illustrating various applications and considerations for successful implementation.
Introduction to Data Mining
Introduction to Data Mining Sushil Kulkarni The document is an introduction to data mining, detailing its goals, processes, and functions, with a focus on the importance of databases. It describes data mining as a method for extracting useful information and patterns through the KDD (knowledge discovery in databases) process, including data selection, preprocessing, transformation, mining, and evaluation. The document further distinguishes between different types of data mining problems, such as classification, clustering, and association rules, underscoring the significance of data mining in real-world applications.
Text MIning
Text MIningPrakhyath Rai This document presents an overview of text mining. It discusses how text mining differs from data mining in that it involves natural language processing of unstructured or semi-structured text data rather than structured numeric data. The key steps of text mining include pre-processing text, applying techniques like summarization, classification, clustering and information extraction, and analyzing the results. Some common applications of text mining are market trend analysis and filtering of spam emails. While text mining allows extraction of information from diverse sources, it requires initial learning systems and suitable programs for knowledge discovery.
Lecture 1: What is Machine Learning?
Lecture 1: What is Machine Learning?Marina Santini The document discusses machine learning, defined as the study of algorithms that learn from data and experience, applicable in various fields like medicine and advertising. It contrasts rule-based systems with machine learning-based systems, emphasizing the importance of generalization and the ability to identify patterns in data. The text covers various types of learning, key concepts, and the relationship between machine learning, statistics, and data mining.
Hadoop Map Reduce
Hadoop Map ReduceVNIT-ACM Student Chapter Hadoop MapReduce is an open source framework for distributed processing of large datasets across clusters of computers. It allows parallel processing of large datasets by dividing the work across nodes. The framework handles scheduling, fault tolerance, and distribution of work. MapReduce consists of two main phases - the map phase where the data is processed key-value pairs and the reduce phase where the outputs of the map phase are aggregated together. It provides an easy programming model for developers to write distributed applications for large scale processing of structured and unstructured data.
Cyber security and AI
Cyber security and AIDexterJanPineda The document discusses cybersecurity, artificial intelligence, and how AI can help improve cybersecurity. It notes that while organizations spend billions on cybersecurity, chief information security officers still feel highly exposed. Traditional security methods focus on preventing infiltration but are always one step behind evolving threats. The document argues that AI can help enforce cyber hygiene practices like least privilege to shrink the attack surface, making the problem more bounded and manageable compared to always chasing threats. It discusses how AI is well-suited for understanding intended application behavior based on established rules and data from good software.
Big data unit i
Big data unit iNavjot Kaur This document provides an overview of big data in various industries. It begins by defining big data and explaining the three V's of big data - volume, variety, and velocity. It then discusses examples of big data in digital marketing, financial services, and healthcare. For digital marketing, it discusses database marketers as pioneers of big data and how big data is transforming digital marketing. For financial services, it discusses how big data is used for fraud detection and credit risk management. It also provides details on algorithmic trading and how it crunches complex interrelated big data. Overall, the document outlines how big data is being leveraged across industries to improve operations, increase revenues, and achieve competitive advantages.
Deep learning - A Visual Introduction
Deep learning - A Visual IntroductionLukas Masuch The document provides an extensive overview of deep learning, a subset of machine learning that excels in recognizing patterns through neural networks modeled after the human brain. It covers the fundamentals of machine learning techniques, algorithms, applications across various domains such as speech and image recognition, as well as the evolution and future prospects of deep learning technology. Key advancements, challenges, and prominent figures in the field are also highlighted, showcasing deep learning's potential impact on society and technology.
Big data lecture notes
Big data lecture notesMohit Saini Big data is high-volume, high-velocity, and high-variety data that is difficult to process using traditional data management tools. It is characterized by 3Vs: volume of data is growing exponentially, velocity as data streams in real-time, and variety as data comes from many different sources and formats. The document discusses big data analytics techniques to gain insights from large and complex datasets and provides examples of big data sources and applications.
Ensemble learning
Ensemble learningHaris Jamil Ensemble learning is a technique that combines multiple models to produce more accurate predictions in both regression and classification tasks. Methods of ensemble learning include voting, averaging, stacking, bagging, and boosting, each employing various algorithms and techniques to improve the performance of the base models. Notably, algorithms like Random Forest and Adaboost are examples of methods used in bagging and boosting, respectively.
Text Mining
Text MiningBiniam Asnake Text mining is the process of extracting useful information and patterns from large collections of unstructured documents. It involves preprocessing texts, applying techniques like categorization, clustering, and summarization, and presenting or visualizing the results. While text mining has many applications in business, science, and other domains, it also faces challenges related to linguistics, analytics, and integrating domain knowledge. The document outlines the definition, techniques, applications, advantages, and limitations of text mining.
Machine Learning
Machine LearningShrey Malik Active learning is a machine learning technique where the learner is able to interactively query the oracle (e.g. a human) to obtain labels for new data points in an effort to learn more accurately from fewer labeled examples. The learner selects the most informative samples to be labeled by the oracle, such as samples closest to the decision boundary or where models disagree most. This allows the learner to minimize the number of labeled samples needed, thus reducing the cost of training an accurate model. Suggested improvements include querying batches of samples instead of single samples and accounting for varying labeling costs.
imageprocessing-abstract
imageprocessing-abstractJagadeesh Kumar The document discusses image processing and describes its goals, applications, and system requirements. It defines image processing as altering existing images in a desired manner to extract important features and provide machine understanding. It provides examples of image processing applications like remote sensing, medical imaging, and character recognition. The proposed system allows users to modify images through tools for compression, rotation, resizing pixels and edge detection, and can process various file formats. Hardware requirements include at least 80GB storage, 512MB RAM, and a Pentium processor, while software requirements include Windows OS, Java/Swing technologies, Apache Tomcat server, and an Oracle or Access backend database.
Machine learning
Machine learningRajib Kumar De Machine learning is a method of data analysis that uses algorithms to iteratively learn from data without being explicitly programmed. It allows computers to find hidden insights in data and become better at tasks via experience. Machine learning has many practical applications and is important due to growing data availability, cheaper and more powerful computation, and affordable storage. It is used in fields like finance, healthcare, marketing and transportation. The main approaches are supervised learning, unsupervised learning, semi-supervised learning, and reinforcement learning. Each has real-world examples like loan prediction, market basket analysis, webpage classification, and marketing campaign optimization.
Big data Presentation
Big data PresentationAswadmehar The document discusses the concept of big data, its characteristics, sources, and implications for various fields including IT and analytics. It highlights the growth and vastness of big data, the need for specialized tools and techniques to manage it, and the potential benefits and risks associated with its application. Furthermore, it emphasizes the necessity for organizations to adapt to the challenges and opportunities presented by big data to remain competitive.
Future of AI - 2023 07 25.pptx
Future of AI - 2023 07 25.pptxGreg Makowski Greg Makowski's presentation discusses the evolution of AI and generative models, highlighting the transition from traditional AI to modern advancements in large language models (LLMs). He covers various AI algorithms, the significance of emergent properties in complex models, and the importance of diverse training data to mitigate biases. The presentation also provides resources and references for further exploration of AI topics and applications.
Big_data_ppt
Big_data_ppt Sadhana Singh The document provides an overview of big data, its evolution, characteristics, and its comparison with traditional RDBMS. It highlights the Hadoop framework, which facilitates the distributed processing of large datasets, covering its components like HDFS and MapReduce. Additionally, it discusses the challenges and applications of big data across various sectors such as retail, travel, and military analytics.
AI, Machine Learning and Deep Learning - The Overview
AI, Machine Learning and Deep Learning - The OverviewSpotle.ai The document provides an overview of artificial intelligence (AI), machine learning (ML), and deep learning (DL), explaining their definitions, types, and applications. It details the progression from narrow AI to the hypothetical superintelligence, outlining the distinct processes and algorithms involved in machine learning and the hierarchical structure of deep learning. Additionally, it discusses real-world applications, such as digital assistants and recommendation systems, highlighting the role of data and algorithms in their functionality.
Big data analytics
Big data analyticsVikram Nandini The document discusses big data analytics. It begins by defining big data as large datasets that are difficult to capture, store, manage and analyze using traditional database management tools. It notes that big data is characterized by the three V's - volume, variety and velocity. The document then covers topics such as unstructured data, trends in data storage, and examples of big data in industries like digital marketing, finance and healthcare.
Main principles of Data Science and Machine Learning
Main principles of Data Science and Machine LearningNikolay Karelin The document provides an overview of machine learning, deep learning, and big data, covering essential concepts, principles, and methodologies. It discusses types of machine learning, risks and challenges, and various datasets sources, along with practical materials for further learning. Key speaker Nikolay Karelin shares insights and resources for both academic and practical applications in AI.
Getting started with Machine Learning
Getting started with Machine LearningGaurav Bhalotia The document provides an overview of machine learning, including definitions, types of machine learning algorithms, and the machine learning process. It defines machine learning as using algorithms to learn from data and make predictions. The main types discussed are supervised learning (classification, regression), unsupervised learning (clustering, association rules), and deep learning using neural networks. The machine learning process involves gathering data, feature engineering, splitting data into training/test sets, selecting an algorithm, training a model, validating it on a validation set, and testing it on a held-out test set. Key enablers of machine learning like large datasets and computing power are also mentioned.
More Related Content
What's hot (20)
Deep Learning Explained
Deep Learning ExplainedMelanie Swan This document summarizes Melanie Swan's presentation on deep learning. It began with defining key deep learning concepts and techniques, including neural networks, supervised vs. unsupervised learning, and convolutional neural networks. It then explained how deep learning works by using multiple processing layers to extract higher-level features from data and make predictions. Deep learning has various applications like image recognition and speech recognition. The presentation concluded by discussing how deep learning is inspired by concepts from physics and statistical mechanics.
ML Basics
ML BasicsSrujanaMerugu1 The document provides a comprehensive overview of machine learning (ML), detailing its applications, methodologies, and the development lifecycle. It emphasizes the importance of correctly framing business problems for ML solutions, selecting appropriate target variables, and understanding data integration and evaluation metrics. Additionally, the document discusses the benefits and challenges of using ML in real-world scenarios, illustrating various applications and considerations for successful implementation.
Introduction to Data Mining
Introduction to Data Mining Sushil Kulkarni The document is an introduction to data mining, detailing its goals, processes, and functions, with a focus on the importance of databases. It describes data mining as a method for extracting useful information and patterns through the KDD (knowledge discovery in databases) process, including data selection, preprocessing, transformation, mining, and evaluation. The document further distinguishes between different types of data mining problems, such as classification, clustering, and association rules, underscoring the significance of data mining in real-world applications.
Text MIning
Text MIningPrakhyath Rai This document presents an overview of text mining. It discusses how text mining differs from data mining in that it involves natural language processing of unstructured or semi-structured text data rather than structured numeric data. The key steps of text mining include pre-processing text, applying techniques like summarization, classification, clustering and information extraction, and analyzing the results. Some common applications of text mining are market trend analysis and filtering of spam emails. While text mining allows extraction of information from diverse sources, it requires initial learning systems and suitable programs for knowledge discovery.
Lecture 1: What is Machine Learning?
Lecture 1: What is Machine Learning?Marina Santini The document discusses machine learning, defined as the study of algorithms that learn from data and experience, applicable in various fields like medicine and advertising. It contrasts rule-based systems with machine learning-based systems, emphasizing the importance of generalization and the ability to identify patterns in data. The text covers various types of learning, key concepts, and the relationship between machine learning, statistics, and data mining.
Hadoop Map Reduce
Hadoop Map ReduceVNIT-ACM Student Chapter Hadoop MapReduce is an open source framework for distributed processing of large datasets across clusters of computers. It allows parallel processing of large datasets by dividing the work across nodes. The framework handles scheduling, fault tolerance, and distribution of work. MapReduce consists of two main phases - the map phase where the data is processed key-value pairs and the reduce phase where the outputs of the map phase are aggregated together. It provides an easy programming model for developers to write distributed applications for large scale processing of structured and unstructured data.
Cyber security and AI
Cyber security and AIDexterJanPineda The document discusses cybersecurity, artificial intelligence, and how AI can help improve cybersecurity. It notes that while organizations spend billions on cybersecurity, chief information security officers still feel highly exposed. Traditional security methods focus on preventing infiltration but are always one step behind evolving threats. The document argues that AI can help enforce cyber hygiene practices like least privilege to shrink the attack surface, making the problem more bounded and manageable compared to always chasing threats. It discusses how AI is well-suited for understanding intended application behavior based on established rules and data from good software.
Big data unit i
Big data unit iNavjot Kaur This document provides an overview of big data in various industries. It begins by defining big data and explaining the three V's of big data - volume, variety, and velocity. It then discusses examples of big data in digital marketing, financial services, and healthcare. For digital marketing, it discusses database marketers as pioneers of big data and how big data is transforming digital marketing. For financial services, it discusses how big data is used for fraud detection and credit risk management. It also provides details on algorithmic trading and how it crunches complex interrelated big data. Overall, the document outlines how big data is being leveraged across industries to improve operations, increase revenues, and achieve competitive advantages.
Deep learning - A Visual Introduction
Deep learning - A Visual IntroductionLukas Masuch The document provides an extensive overview of deep learning, a subset of machine learning that excels in recognizing patterns through neural networks modeled after the human brain. It covers the fundamentals of machine learning techniques, algorithms, applications across various domains such as speech and image recognition, as well as the evolution and future prospects of deep learning technology. Key advancements, challenges, and prominent figures in the field are also highlighted, showcasing deep learning's potential impact on society and technology.
Big data lecture notes
Big data lecture notesMohit Saini Big data is high-volume, high-velocity, and high-variety data that is difficult to process using traditional data management tools. It is characterized by 3Vs: volume of data is growing exponentially, velocity as data streams in real-time, and variety as data comes from many different sources and formats. The document discusses big data analytics techniques to gain insights from large and complex datasets and provides examples of big data sources and applications.
Ensemble learning
Ensemble learningHaris Jamil Ensemble learning is a technique that combines multiple models to produce more accurate predictions in both regression and classification tasks. Methods of ensemble learning include voting, averaging, stacking, bagging, and boosting, each employing various algorithms and techniques to improve the performance of the base models. Notably, algorithms like Random Forest and Adaboost are examples of methods used in bagging and boosting, respectively.
Text Mining
Text MiningBiniam Asnake Text mining is the process of extracting useful information and patterns from large collections of unstructured documents. It involves preprocessing texts, applying techniques like categorization, clustering, and summarization, and presenting or visualizing the results. While text mining has many applications in business, science, and other domains, it also faces challenges related to linguistics, analytics, and integrating domain knowledge. The document outlines the definition, techniques, applications, advantages, and limitations of text mining.
Machine Learning
Machine LearningShrey Malik Active learning is a machine learning technique where the learner is able to interactively query the oracle (e.g. a human) to obtain labels for new data points in an effort to learn more accurately from fewer labeled examples. The learner selects the most informative samples to be labeled by the oracle, such as samples closest to the decision boundary or where models disagree most. This allows the learner to minimize the number of labeled samples needed, thus reducing the cost of training an accurate model. Suggested improvements include querying batches of samples instead of single samples and accounting for varying labeling costs.
imageprocessing-abstract
imageprocessing-abstractJagadeesh Kumar The document discusses image processing and describes its goals, applications, and system requirements. It defines image processing as altering existing images in a desired manner to extract important features and provide machine understanding. It provides examples of image processing applications like remote sensing, medical imaging, and character recognition. The proposed system allows users to modify images through tools for compression, rotation, resizing pixels and edge detection, and can process various file formats. Hardware requirements include at least 80GB storage, 512MB RAM, and a Pentium processor, while software requirements include Windows OS, Java/Swing technologies, Apache Tomcat server, and an Oracle or Access backend database.
Machine learning
Machine learningRajib Kumar De Machine learning is a method of data analysis that uses algorithms to iteratively learn from data without being explicitly programmed. It allows computers to find hidden insights in data and become better at tasks via experience. Machine learning has many practical applications and is important due to growing data availability, cheaper and more powerful computation, and affordable storage. It is used in fields like finance, healthcare, marketing and transportation. The main approaches are supervised learning, unsupervised learning, semi-supervised learning, and reinforcement learning. Each has real-world examples like loan prediction, market basket analysis, webpage classification, and marketing campaign optimization.
Big data Presentation
Big data PresentationAswadmehar The document discusses the concept of big data, its characteristics, sources, and implications for various fields including IT and analytics. It highlights the growth and vastness of big data, the need for specialized tools and techniques to manage it, and the potential benefits and risks associated with its application. Furthermore, it emphasizes the necessity for organizations to adapt to the challenges and opportunities presented by big data to remain competitive.
Future of AI - 2023 07 25.pptx
Future of AI - 2023 07 25.pptxGreg Makowski Greg Makowski's presentation discusses the evolution of AI and generative models, highlighting the transition from traditional AI to modern advancements in large language models (LLMs). He covers various AI algorithms, the significance of emergent properties in complex models, and the importance of diverse training data to mitigate biases. The presentation also provides resources and references for further exploration of AI topics and applications.
Big_data_ppt
Big_data_ppt Sadhana Singh The document provides an overview of big data, its evolution, characteristics, and its comparison with traditional RDBMS. It highlights the Hadoop framework, which facilitates the distributed processing of large datasets, covering its components like HDFS and MapReduce. Additionally, it discusses the challenges and applications of big data across various sectors such as retail, travel, and military analytics.
AI, Machine Learning and Deep Learning - The Overview
AI, Machine Learning and Deep Learning - The OverviewSpotle.ai The document provides an overview of artificial intelligence (AI), machine learning (ML), and deep learning (DL), explaining their definitions, types, and applications. It details the progression from narrow AI to the hypothetical superintelligence, outlining the distinct processes and algorithms involved in machine learning and the hierarchical structure of deep learning. Additionally, it discusses real-world applications, such as digital assistants and recommendation systems, highlighting the role of data and algorithms in their functionality.
Big data analytics
Big data analyticsVikram Nandini The document discusses big data analytics. It begins by defining big data as large datasets that are difficult to capture, store, manage and analyze using traditional database management tools. It notes that big data is characterized by the three V's - volume, variety and velocity. The document then covers topics such as unstructured data, trends in data storage, and examples of big data in industries like digital marketing, finance and healthcare.
Similar to Machine Learning Project Lifecycle (20)
Main principles of Data Science and Machine Learning
Main principles of Data Science and Machine LearningNikolay Karelin The document provides an overview of machine learning, deep learning, and big data, covering essential concepts, principles, and methodologies. It discusses types of machine learning, risks and challenges, and various datasets sources, along with practical materials for further learning. Key speaker Nikolay Karelin shares insights and resources for both academic and practical applications in AI.
Getting started with Machine Learning
Getting started with Machine LearningGaurav Bhalotia The document provides an overview of machine learning, including definitions, types of machine learning algorithms, and the machine learning process. It defines machine learning as using algorithms to learn from data and make predictions. The main types discussed are supervised learning (classification, regression), unsupervised learning (clustering, association rules), and deep learning using neural networks. The machine learning process involves gathering data, feature engineering, splitting data into training/test sets, selecting an algorithm, training a model, validating it on a validation set, and testing it on a held-out test set. Key enablers of machine learning like large datasets and computing power are also mentioned.
ML_big_picture-2.0.pptx
ML_big_picture-2.0.pptxFrancis Piéraut This document provides an overview of machine learning and Francis Pieraut's experience with it. It discusses the evolution of machine learning from expert systems to modern techniques using deep learning. It also addresses key aspects of machine learning like the importance of feature engineering, regularization techniques to prevent overfitting, and how the cost function and large datasets are crucial. Throughout, it emphasizes that machine learning requires careful data preparation and domain expertise to apply effectively.
Machine learning workshop
Machine learning workshopLakshya Sivaramakrishnan The document introduces machine learning and TensorFlow, explaining concepts such as supervised, unsupervised, and reinforcement learning, along with practical use cases of machine learning. It highlights the importance of algorithms in learning patterns from data and discusses how to get started with machine learning using existing models or cloud-based APIs. Additionally, it outlines the TensorFlow library's capabilities and provides code examples for basic operations and model training with the MNIST dataset.
Deep learning introduction
Deep learning introductionAdwait Bhave The document discusses the history and evolution of deep learning, highlighting its relevance in artificial intelligence, machine learning, and various applications across multiple industries such as healthcare, finance, and autonomous driving. It explains key concepts including neural networks, backpropagation, and feature extraction, placing emphasis on the importance of large datasets and model tuning for effective algorithm performance. The potential future impacts and challenges of deep learning are also addressed, alongside examples of successful implementations in real-world scenarios.
Applying Machine Learning and Artificial Intelligence to Business
Applying Machine Learning and Artificial Intelligence to BusinessRussell Miles The document discusses the applications, algorithms, and advancements in machine learning (ML) and artificial intelligence (AI), emphasizing their importance in various business and technological contexts. It details types of learning, various ML and deep learning toolkits, notable AI companies, and milestones in AI history. The document also explores the opportunities and threats presented by AI and robotics in modern society.
Practical ML
Practical MLAntonio Pitasi - The document provides an introduction to machine learning concepts and practical examples using neural networks. It discusses different machine learning categories and algorithms. It then demonstrates how to build and train simple feedforward neural networks to classify points and recognize handwritten digits. Code examples are provided using Python libraries like Keras. References are included for further reading.
S3_ML Introduction.pdf
S3_ML Introduction.pdfSumanshuVerma Machine learning is a subset of artificial intelligence that uses algorithms and data to gradually improve accuracy. There are several types of machine learning including supervised learning, which uses labeled training data; unsupervised learning, which finds patterns without labels; and reinforcement learning, which uses trial-and-error. Machine learning has many applications such as image classification, customer segmentation, and fraud detection. It provides benefits like accuracy, automation, and scalability but also has limitations such as potential bias if the training data is biased.
10 more lessons learned from building Machine Learning systems - MLConf
10 more lessons learned from building Machine Learning systems - MLConfXavier Amatriain 1. Machine learning applications at Quora include answer ranking, feed ranking, topic recommendations, user recommendations, and more. A variety of models are used including logistic regression, gradient boosted decision trees, neural networks, and matrix factorization.
2. Implicit signals like watching and clicking tend to be more useful than explicit signals like ratings. However, both implicit and explicit signals combined can better represent long-term goals.
3. It is important to focus on feature engineering to create features that are reusable, transformable, interpretable, and reliable. The outputs of models may become inputs to other models, so care must be taken to avoid feedback loops and ensure proper data dependencies.
Xavier Amatriain, VP of Engineering, Quora at MLconf SF - 11/13/15
Xavier Amatriain, VP of Engineering, Quora at MLconf SF - 11/13/15MLconf The document outlines ten lessons learned from building machine learning (ML) systems, focusing on the importance of implicit signals over explicit ones, the necessity of proper training data, and the effectiveness of using ensembles. Key insights emphasize the integration of supervised and unsupervised learning, the significance of robust feature engineering, and the need for a dual-focus ML infrastructure for experimentation and production. Additionally, it discusses the roles of data scientists and ML engineers in the innovation process and stresses the importance of model debuggability.
10 more lessons learned from building Machine Learning systems
10 more lessons learned from building Machine Learning systemsXavier Amatriain 1. Machine learning applications at Quora include answer ranking, feed ranking, topic recommendations, user recommendations, and more. A variety of models are used including logistic regression, gradient boosted decision trees, neural networks, and matrix factorization.
2. Implicit signals like watching and clicking tend to be more useful than explicit signals like ratings. However, both implicit and explicit signals combined can better represent long-term goals.
3. The outputs of machine learning models will often become inputs to other models, so models need to be designed with this in mind to avoid issues like feedback loops.
MLSEV Virtual. ML Platformization and AutoML in the Enterprise
MLSEV Virtual. ML Platformization and AutoML in the EnterpriseBigML, Inc The document discusses the importance of machine learning (ML) platforms for adopting ML at scale within enterprises, emphasizing their role in automating and optimizing business processes. It outlines various ML platforms created by companies such as Facebook, Uber, and Airbnb, comparing custom-built solutions versus commercial offerings. Additionally, it highlights the current challenges, including technical debt and the need for organizations to transition from AI exploration to sustainable AI systems.
MachinaFiesta: A Vision into Machine Learning 🚀
MachinaFiesta: A Vision into Machine Learning 🚀GDSCNiT The document discusses the fundamentals of artificial intelligence (AI), machine learning (ML), and deep learning (DL) with a focus on their interrelationship and learning methodologies. It highlights the significance of datasets in training AI models, categorizes various learning methods, and introduces neural networks and tensor structures. Additionally, it covers the emerging field of generative AI, detailing its capabilities, market trends, production challenges, and potential solutions.
Ml masterclass
Ml masterclassMaxwell Rebo This document provides an overview of machine learning concepts and example algorithms. It discusses how machine learning systems can learn from experience without explicit programming. It then covers classification and regression problems and provides examples of random forests and Gaussian processes algorithms. The document also discusses feature learning with examples of autoencoders and PCA. Finally, it discusses practical considerations for applying machine learning, including the importance of data quality, data pipelines, managing error risk, and institutionalizing machine learning applications.
Implementing Machine Learning in the Real World
Implementing Machine Learning in the Real WorldJesus Rodriguez This document outlines 15 lessons learned from building large-scale machine learning systems in the real world. Some key challenges discussed include data scientists not being well-suited for engineering work, traditional development methodologies not working for machine learning, the difficulty of data labeling and feature extraction, and the complexities of training, executing, operationalizing, and securing machine learning models at scale. The document provides ideas to address these challenges such as establishing separate data science and engineering teams, implementing automated data labeling strategies, leveraging centralized feature stores, and adopting techniques like transfer learning and continual learning.
AI Ml Introduction with images and examples.pptx
AI Ml Introduction with images and examples.pptxbajajrishabh96tech This document provides an overview of artificial intelligence (AI) and machine learning (ML). It defines AI as the ability of computers to mimic human intelligence through tasks like problem solving and learning. ML is a subset of AI that allows systems to learn from data without being explicitly programmed. The document outlines key concepts in AI and ML like perception, reasoning, and supervised/unsupervised learning. It also discusses applications in healthcare, finance, autonomous vehicles and more. Popular ML techniques like recommender systems, fraud detection, and image recognition are covered. The presentation concludes by exploring challenges, future trends, and the growing role of deep learning.
Machine Learning Teams - Full Stack Deep Learning
Machine Learning Teams - Full Stack Deep LearningSergey Karayev The document discusses the complexities of managing machine learning (ML) teams, emphasizing the need for effective hiring, organizational structure, and project management methodologies. It outlines various roles within ML teams, such as ML product managers, engineers, and researchers, along with the necessary skills for each role. Additionally, it presents best practices for structuring ML organizations based on maturity levels and highlights the challenges in communicating and aligning expectations within technical and leadership domains.
GECBSP Guide GDSC India ML Study Jam organizer guide modified.pdf
GECBSP Guide GDSC India ML Study Jam organizer guide modified.pdfDomendra Sahu This document provides information about an ML Study Jams program from April 18th to May 30th 2022. It discusses machine learning trends, jobs, and salaries in the field. It also introduces common ML approaches like supervised, unsupervised, reinforcement, and semi-supervised learning. The 6-week program covers introductory and intermediate ML topics through online sessions and assignments. Participants can get certified in TensorFlow and are eligible for reimbursement. The goal is to upskill community members and encourage peer learning through hands-on projects and a career panel discussion.
Computer Design Concepts for Machine Learning
Computer Design Concepts for Machine LearningFacultad de Informática UCM The document discusses the intersection of computer design and machine learning, emphasizing the evolution of AI, neural networks, and deep learning. It highlights the challenges and advancements in hardware and algorithms, particularly in the context of architectures like GPUs, TPUs, and FPGAs for optimizing computations. Key issues such as data quality, explainability, and biases in AI are also addressed, underscoring the complexity and potential of machine learning technologies.
MLSD18. End-to-End Machine Learning
MLSD18. End-to-End Machine LearningBigML, Inc The document summarizes a presentation about machine learning. It discusses what machine learning is and is not, provides examples of machine learning tasks like predictive maintenance and personalized music, and addresses realities of applying machine learning like the fact that all models are wrong and the concepts of overfitting and underfitting. It also discusses choosing machine learning algorithms and automating the process.
Ad
Recently uploaded (20)
10 Key Challenges for AI within the EU Data Protection Framework.pdf
10 Key Challenges for AI within the EU Data Protection Framework.pdfPriyanka Aash 10 Key Challenges for AI within the EU Data Protection Framework
The Future of Product Management in AI ERA.pdf
The Future of Product Management in AI ERA.pdfAlyona Owens Hi, I’m Aly Owens, I have a special pleasure to stand here as over a decade ago I graduated from CityU as an international student with an MBA program. I enjoyed the diversity of the school, ability to work and study, the network that came with being here, and of course the price tag for students here has always been more affordable than most around.
Since then I have worked for major corporations like T-Mobile and Microsoft and many more, and I have founded a startup. I've also been teaching product management to ensure my students save time and money to get to the same level as me faster avoiding popular mistakes. Today as I’ve transitioned to teaching and focusing on the startup, I hear everybody being concerned about Ai stealing their jobs… We’ll talk about it shortly.
But before that, I want to take you back to 1997. One of my favorite movies is “Fifth Element”. It wowed me with futuristic predictions when I was a kid and I’m impressed by the number of these predictions that have already come true. Self-driving cars, video calls and smart TV, personalized ads and identity scanning. Sci-fi movies and books gave us many ideas and some are being implemented as we speak. But we often get ahead of ourselves:
Flying cars,Colonized planets, Human-like AI: not yet, Time travel, Mind-machine neural interfaces for everyone: Only in experimental stages (e.g. Neuralink).
Cyberpunk dystopias: Some vibes (neon signs + inequality + surveillance), but not total dystopia (thankfully).
On the bright side, we predict that the working hours should drop as Ai becomes our helper and there shouldn’t be a need to work 8 hours/day. Nobody knows for sure but we can require that from legislation. Instead of waiting to see what the government and billionaires come up with, I say we should design our own future.
So, we as humans, when we don’t know something - fear takes over. The same thing happened during the industrial revolution. In the Industrial Era, machines didn’t steal jobs—they transformed them but people were scared about their jobs. The AI era is making similar changes except it feels like robots will take the center stage instead of a human. First off, even when it comes to the hottest space in the military - drones, Ai does a fraction of work. AI algorithms enable real-time decision-making, obstacle avoidance, and mission optimization making drones far more autonomous and capable than traditional remote-controlled aircraft. Key technologies include computer vision for object detection, GPS-enhanced navigation, and neural networks for learning and adaptation. But guess what? There are only 2 companies right now that utilize Ai in drones to make autonomous decisions - Skydio and DJI.
A Constitutional Quagmire - Ethical Minefields of AI, Cyber, and Privacy.pdf
A Constitutional Quagmire - Ethical Minefields of AI, Cyber, and Privacy.pdfPriyanka Aash A Constitutional Quagmire - Ethical Minefields of AI, Cyber, and Privacy
“MPU+: A Transformative Solution for Next-Gen AI at the Edge,” a Presentation...
“MPU+: A Transformative Solution for Next-Gen AI at the Edge,” a Presentation...Edge AI and Vision Alliance For the full video of this presentation, please visit: https://www.edge-ai-vision.com/2025/06/mpu-a-transformative-solution-for-next-gen-ai-at-the-edge-a-presentation-from-fotonation/
Petronel Bigioi, CEO of FotoNation, presents the “MPU+: A Transformative Solution for Next-Gen AI at the Edge” tutorial at the May 2025 Embedded Vision Summit.
In this talk, Bigioi introduces MPU+, a novel programmable, customizable low-power platform for real-time, localized intelligence at the edge. The platform includes an AI-augmented image signal processor that enables leading image and video quality.
In addition, it integrates ultra-low-power object and motion detection capabilities to enable always-on computer vision. A programmable neural processor provides flexibility to efficiently implement new neural networks. And additional specialized engines facilitate image stabilization and audio enhancements.
Enhance GitHub Copilot using MCP - Enterprise version.pdf
Enhance GitHub Copilot using MCP - Enterprise version.pdfNilesh Gule Slide deck related to the GitHub Copilot Bootcamp in Melbourne on 17 June 2025
AI Agents and FME: A How-to Guide on Generating Synthetic Metadata
AI Agents and FME: A How-to Guide on Generating Synthetic MetadataSafe Software In the world of AI agents, semantics is king. Good metadata is thus essential in an organization's AI readiness checklist. But how do we keep up with the massive influx of new data? In this talk we go over the tips and tricks in generating synthetic metadata for the consumption of human users and AI agents alike.
PyCon SG 25 - Firecracker Made Easy with Python.pdf
PyCon SG 25 - Firecracker Made Easy with Python.pdfMuhammad Yuga Nugraha Explore the ease of managing Firecracker microVM with the firecracker-python. In this session, I will introduce the basics of Firecracker microVM and demonstrate how this custom SDK facilitates microVM operations easily. We will delve into the design and development process behind the SDK, providing a behind-the-scenes look at its creation and features. While traditional Firecracker SDKs were primarily available in Go, this module brings a simplicity of Python to the table.
"How to survive Black Friday: preparing e-commerce for a peak season", Yurii ...
"How to survive Black Friday: preparing e-commerce for a peak season", Yurii ...Fwdays We will explore how e-commerce projects prepare for the busiest time of the year, which key aspects to focus on, and what to expect. We’ll share our experience in setting up auto-scaling, load balancing, and discuss the loads that Silpo handles, as well as the solutions that help us navigate this season without failures.
The Growing Value and Application of FME & GenAI
The Growing Value and Application of FME & GenAISafe Software With the cost of using Generative AI services dropping exponentially and the array of available models continually expanding, integrating AI into FME workflows has become inexpensive, accessible and effective. This presentation explores how GenAI within FME can cost-effectively transform data workflows by automating data extraction, validation, classification and augmentation tasks. We’ll discuss how FME’s no-code flexibility enables users to combine Generative AI and Computer Vision tools that create efficient workflows tailored to specific challenges. Using recent practical examples, we’ll demonstrate how these integrations can simplify complex tasks, save time and enhance data quality.
9-1-1 Addressing: End-to-End Automation Using FME
9-1-1 Addressing: End-to-End Automation Using FMESafe Software This session will cover a common use case for local and state/provincial governments who create and/or maintain their 9-1-1 addressing data, particularly address points and road centerlines. In this session, you'll learn how FME has helped Shelby County 9-1-1 (TN) automate the 9-1-1 addressing process; including automatically assigning attributes from disparate sources, on-the-fly QAQC of said data, and reporting. The FME logic that this presentation will cover includes: Table joins using attributes and geometry, Looping in custom transformers, Working with lists and Change detection.
Securing Account Lifecycles in the Age of Deepfakes.pptx
Securing Account Lifecycles in the Age of Deepfakes.pptxFIDO Alliance Securing Account Lifecycles in the Age of Deepfakes
Security Tips for Enterprise Azure Solutions
Security Tips for Enterprise Azure SolutionsMichele Leroux Bustamante Delivering solutions to Azure may involve a variety of architecture patterns involving your applications, APIs data and associated Azure resources that comprise the solution. This session will use reference architectures to illustrate the security considerations to protect your Azure resources and data, how to achieve Zero Trust, and why it matters. Topics covered will include specific security recommendations for types Azure resources and related network security practices. The goal is to give you a breadth of understanding as to typical security requirements to meet compliance and security controls in an enterprise solution.
Coordinated Disclosure for ML - What's Different and What's the Same.pdf
Coordinated Disclosure for ML - What's Different and What's the Same.pdfPriyanka Aash Coordinated Disclosure for ML - What's Different and What's the Same
WebdriverIO & JavaScript: The Perfect Duo for Web Automation
WebdriverIO & JavaScript: The Perfect Duo for Web Automationdigitaljignect In today’s dynamic digital landscape, ensuring the quality and dependability of web applications is essential. While Selenium has been a longstanding solution for automating browser tasks, the integration of WebdriverIO (WDIO) with Selenium and JavaScript marks a significant advancement in automation testing. WDIO enhances the testing process by offering a robust interface that improves test creation, execution, and management. This amalgamation capitalizes on the strengths of both tools, leveraging Selenium’s broad browser support and WDIO’s modern, efficient approach to test automation. As automation testing becomes increasingly vital for faster development cycles and superior software releases, WDIO emerges as a versatile framework, particularly potent when paired with JavaScript, making it a preferred choice for contemporary testing teams.
Salesforce Summer '25 Release Frenchgathering.pptx.pdf
Salesforce Summer '25 Release Frenchgathering.pptx.pdfyosra Saidani Salesforce Summer '25 Release Frenchgathering.pptx.pdf
Lessons Learned from Developing Secure AI Workflows.pdf
Lessons Learned from Developing Secure AI Workflows.pdfPriyanka Aash Lessons Learned from Developing Secure AI Workflows
GenAI Opportunities and Challenges - Where 370 Enterprises Are Focusing Now.pdf
GenAI Opportunities and Challenges - Where 370 Enterprises Are Focusing Now.pdfPriyanka Aash GenAI Opportunities and Challenges - Where 370 Enterprises Are Focusing Now
EIS-Webinar-Engineering-Retail-Infrastructure-06-16-2025.pdf
EIS-Webinar-Engineering-Retail-Infrastructure-06-16-2025.pdfEarley Information Science As AI reshapes expectations in retail and B2B commerce, organizations are recognizing a critical reality: meaningful AI outcomes depend on well-structured, adaptable infrastructure. In this session, Seth Earley is joined by Phil Ryan - AI strategist, search technologist, and founder of Glass Leopard Technologies - for a candid conversation on what it truly means to engineer systems for scale, agility, and intelligence.
Phil draws on more than two decades of experience leading search and AI initiatives for enterprise organizations. Together, he and Seth explore the challenges businesses face when legacy architectures limit personalization, agility, and real-time decisioning - and what needs to change to support agentic technologies and next-best-action capabilities.
Key themes from the webinar include:
Composability as a prerequisite for AI - Why modular, loosely coupled systems are essential for adapting to rapid innovation and evolving business needs
Search and relevance as foundational to AI - How techniques honed-in enterprise search have laid the groundwork for more responsive and intelligent customer experiences
From MDM and CDP to agentic systems - How data platforms are evolving to support richer customer context and dynamic orchestration
Engineering for business alignment - Why successful AI programs require architectural decisions grounded in measurable outcomes
The conversation is practical and forward-looking, connecting deep technical understanding with real-world business needs. Whether you’re modernizing your commerce stack or exploring how AI can enhance product discovery, personalization, or customer journeys, this session provides a clear-eyed view of the capabilities, constraints, and priorities that matter most.
“MPU+: A Transformative Solution for Next-Gen AI at the Edge,” a Presentation...
“MPU+: A Transformative Solution for Next-Gen AI at the Edge,” a Presentation...Edge AI and Vision Alliance
Ad
Machine Learning Project Lifecycle
- 2. Content • Machine Learning Everywhere • ML Project Lifecycle • Deep NN Training Focus • How can I Learn? • How can I Apply? • Demos 2
- 4. Use Cases Forbes: The Top 10 AI And Machine Learning Use Cases Everyone Should Know About 1. Data Security, 2. Personal Security, 3. Financial Trading, 4. Healthcare, 5. Marketing personalization, 6. Fraud Detection, 7. Recommendations, 8. Online Search, 9. NLP, 10. Smart Cars 4
- 5. Internet of Things • 25 and 50 billion Internet-connected devices By 2020, • IoT generates Big Data • Volume • Velocity in terms of time and location dependency, • Variety of multiple modalities • Varying data quality • Nowadays, Train in the Cloud or Fog, predict locally • Tomorrow, Train and predict locally 5
- 7. Products • Text Analysis • uclassify.com • Speech <-> Text Recognition • Google Cloud • IBM Watson • Image and Video Classification • clarifai.com • Etc, etc, etc. 7
- 12. 0. Define • Data • Rows/Example/Instance/Input/ /Observation/Record/Point /Sample/Entity : x(i) • Columns/Feature/Variable/Predictor /Characteristic/Field/Attribute : x(i) j • Quantitative (numeric, continue) • Qualitative (textual, category) • Dimension, Visualization • m Examples: i = 1..m • n Features: j = 1..n • Label/class/output : yi • For each example (0/1) University F1 F2 F3 F4 F5 F6 Region Budget (MD) Insertion (%) Shanghai ranking? U1 979 561 186 786 835 536 N 7 72.00 1 U2 895 247 985 206 870 246 N 7 67.00 1 U3 344 889 643 951 783 162 E 2 35.00 0 U4 400 959 999 312 981 254 W 8 57.00 0 U5 243 521 393 596 400 138 E 6 29.00 1 U6 882 722 518 541 425 551 W 10 80.00 1 U7 814 193 829 192 334 597 E 8 68.00 0 U8 186 972 763 968 217 772 S 1 20.00 0 U9 647 656 527 393 738 813 S 7 65.00 1 U10 568 570 458 799 682 530 N 4 59.00 1 12
- 13. 0. Define • Data • Structured • CSV, XML, JSON, XLSX, etc. • Unstructured • DOC, HTML, PDF, PNG, MP3, MP4, etc. 13
- 14. Image via Abdul Rahid 14 0. Define
- 15. 0. Define x1 x2 x1 x2 Supervised Learning Unsupervised Learning Clustering Reinforcement Learning 15
- 18. 1. Collect • Government, NGOs, Scientific Communities, etc. • Open Data Platforms, • Web data: AWS, Twitter API, Facebook API, Scraping tools, etc. • Mobile Data: google maps, logs, etc • Enterprise data: Databases, Datawerhouses • IoT, Industrial IoT data • More: Forbes: Big Data: 33 Brilliant And Free Data Sources Anyone Can Use 18
- 19. 2. Prepare Dimensionality Reduction Visualization m>>n m>n m<n Feature Extraction Feature Selection 19 Data Unbalance
- 22. 4. Train 22 Training set Validation set Test (Blind) set X Y Training set Validation set Test set Models Training Feature Selection Model Selection (Hypertune) Performance Analysis (Evaluate) Best Model Y Training Inference
- 23. 4. Train 23
- 30. 4. Train 30 ANN 𝑎 * [,] 𝑎 . [,] 𝑎 , [,] 𝑎, [*] 𝑦( 𝑧, , = 𝑤, , 2 𝑥 + 𝑏, [,] , 𝑎 , [,] = 𝑔(𝑧, , ) 𝑧* , = 𝑤* , 2 𝑥 + 𝑏* [,] , 𝑎 * [,] = 𝑔(𝑧* , ) 𝑧. , = 𝑤. , 2 𝑥 + 𝑏. [,] , 𝑎 . [,] = 𝑔(𝑧. , ) 𝑧, * = 𝑤, * 2 𝑥 + 𝑏, [*] , 𝑎 , [*] = 𝑔(𝑧, * ) Regularization term x1 x2 + 𝜆 2𝑚 :: 𝑊< =* > =?, @ <?, min D,E 𝐽(𝑊, 𝑏)
- 32. 5. Evaluate 32 𝐹𝑃𝑅 = 𝐹𝑃 𝑁 𝑇𝑃𝑅 = 𝑇𝑃 𝑃
- 37. Deep NN main architectures 37 𝑥, 𝑥* 𝑥. 𝑦( • Feed Forward NN • One hidden layer • Deep NN • Multiple hidden layer
- 38. Deep NN main architectures 38 𝑦(POOLCONV POOLCONVX • Convolutional NN • Computer Vision • LeNet-5, AlexNet, VGG-16, ResNet, Inception
- 39. Deep NN main architectures 39 𝑥, 𝑥* 𝑥. 𝑦( 𝑥, 𝑥* 𝑥. 𝑦( 𝑥, 𝑥* 𝑥. 𝑦( 𝑥, 𝑥* 𝑥. 𝑦( time • Recurrent NN • Sequential models • NLP, Video, Signal • Problem of vanishing gradient • Gating Units: LSTM, GRU 𝑥, 𝑥* 𝑥. General Recurrent NN
- 40. Deep NN main architectures 40 • Tree Recursive NN • Understand compositionality • NLP parsing, Vision • Tree based • Recursive NN NN NN NN NN
- 41. Deep NN main architectures • Unsupervised Pre-trained NN • No Random Initialization • Parameters are initialized with unsupervised manner • Restricted Botzman Machines • Autoencoders • Regularization effect • Solves Overfitting • Transfer Learning • Model trained on one task is re-purposed on a second related task 41 𝑥, 𝑥* 𝑥. 𝑦( Pre-train Pre-train Pre-trainPre-train Autoencoders
- 42. Activation Functions 42 z Sigmoid 𝑎 = 1 1 + 𝑒NO z 𝑎 = 𝑒O − 𝑒NO 𝑒O + 𝑒NOTanh z ReLU 𝑎 = max (0, 𝑧) z Leaky ReLU 𝑎 = max (0.01𝑧, 𝑧) Rectified Linear Unit z Linear (identity) 𝑎 = 𝑧 z Softmax 𝑎< = 𝑒OU ∑ 𝑒OUW <?, Multi-class Classification Binary classification regression Output Layer Hidden Layers
- 43. 43 Output Layer Input Layer Hidden Layer Activation Functions use linear when y is real (regression) use ReLU when y is real positive (regression) use sigmoid when y in {0,1} (2 classes) use softmax when y in {0,1,2, ...} (K classes) use tanh, ReLU or leaky ReLU 𝑎 * [,] 𝑎 . [,] 𝑎 , [,] 𝑎, [*] 𝑦(
- 44. Optimization (solve underfitting) • Back-Propagation • Gradient Descent • NormalizingInputs to speed up leaning • Gradient checking to prevent Vanishing/Exploding gradient • Momentumto prevent oscillations 44 Learning Rate
- 45. Optimization (solve underfitting) • Gradient Descent (Batch, Mini-batch, Stochastic) • Constant Learning Rate Decay • Time based decay: lr = lr0/(1+kt) • Step decay: lr = lr0 * drop^floor(epoch / epochs_drop) • Exponential decay lr = lr0 * exp(−kt) • Adaptive Learning Rate Decay • Adagrad • Adadelta • RMSprop • Adam • More details: Yoshua Bengio Paper 45
- 49. How can I Learn? • Math • Statistics, Probabilistic Graphical Models, Algebra, Optimization • Programming Languages • Python, R, others • Books • Gareth James, Daniela Witten, Trevor Hastie, Robert Tibshirani. “An introduction to statistical learning with applications in R”. 2013. • Tom M. Mitchell. “Machine Learning”. 1997 • Others 49
- 50. How can I Learn? • MOOCs • Coursera: Machine learning, Deep learning specialization • Udacity, Udemy, Fun, etc. • StackOverflow • Research Papers • Read and rewrite algorithms from scratch • Follow People • Androw Ng, Yann LeCun, Jeff Hinton, Yoshua Bengio, Chris Manning, Sebastian Thrun, etc. 50
- 52. How can I Apply? • Start small projects and use Framworks • TensorFlow, Keras, Caffe, Microsoft Cognitive Toolkit (CNTK 2), MXNet, Scikit-learn, Spark MLlib, etc. • Kaggle, UCI Machine Learning Repo • Find data • Go for competitions • Notebooks: • Jupyter notebook, google cloud Datalab • Github • Find codes • Share your code • Softwares • Knime: open sourcedata analytics • IBM SPSS Modeler: data mining and text analytics software application from IBM 52
- 54. How can I Apply? • Google ML APIs • Vision API • Video Intelligence API • Speech API • Natural Language API • Translation API 54
- 55. Demos 55
- 56. My PhD students demos (Youtube) Younes Choubik Marouane Hamda Marouane Benmoussa Hicham Boudkik 56 Mohamed El Kaddoury Mohamed Es-Salhy
- 57. THANKS! 57