Feature Engineering in H2O Driverless AI - Dmitry Larko - H2O AI World London...Sri Ambati
This talk was recorded in London on October 30th, 2018 and can be viewed here: https://youtu.be/d6UMEmeXB6o
In his talk Dmitry is going to cover common feature engineering techniques used to build robust machine learning models as well as some not widely known/used approaches.
Bio: Senior Data Scientist at H2O.ai, Dmitry Larko also is a former #25 Kaggle Grandmaster and loves to use his machine learning and data science skills in Kaggle Competitions and predictive analytics software development. He has more than 15 years of experience in information technology. Post his masters in computer information systems from Krasnoyarsk State Technical University (KSTU), he started his career in data warehousing and business intelligence and gradually moved to big data and data science. He holds a lot of experience in predictive analytics in a wide array of domains and tasks. Prior to H2O.ai, Dmitry held the position of SAP BW Developer at Chevron, Data Scientist at EPAM, and that of Lead Software Engineer with the Russian Federation.
Feature Engineering for ML - Dmitry Larko, H2O.aiSri Ambati
This talk was given at H2O World 2018 NYC and can be viewed here: https://youtu.be/wcFdmQSX6hM
Description:
In this talk, Dmitry shares his approach to feature engineering which he used successfully in various Kaggle competitions. He covers common techniques used to convert your features into numeric representation used by ML algorithms.
Speaker's Bio:
Dmitry has more than 10 years of experience in IT. Starting with data warehousing and BI, now in big data and data science. He has a lot of experience in predictive analytics software development for different domains and tasks. He is also a Kaggle Grandmaster who loves to use his machine learning and data science skills on Kaggle competitions.
Leakage in Meta Modeling And Its Connection to HCC Target-Encoding - Mathias ...Sri Ambati
Presented at #H2OWorld 2017 in Mountain View, CA.
Enjoy the video: https://youtu.be/Q0QmziFcfU0.
Learn more about H2O.ai: https://www.h2o.ai/.
Follow @h2oai: https://twitter.com/h2oai.
- - -
Abstract:
This talk covers the problems of leakage in a standard k-fold meta modeling (ensembling / stacking) setup, the connection to HCC target encoding and how to properly deal with the induced consequences in order to detect and finally avoid overfitting.
Mathias' Bio:
A Kaggle Grandmaster and a Data Scientist at H2O.ai, Mathias Müller holds an AI and ML focused diploma (eq. M.Sc.) in computer science from Humboldt University in Berlin. During his studies, he keenly worked on computer vision in the context of bio-inspired visual navigation of autonomous flying quadrocopters. Prior to H2O.ai, he as a machine learning engineer for FSD Fahrzeugsystemdaten GmbH in the automotive sector. His stint with Kaggle was a chance encounter as he stumbled upon the data competition platform while looking for a more ML-focused platform as compared to TopCoder. This is where he entered his first predictive modeling competition and climbed up the ladder to be a Grandmaster. He is an active contributor to XGBoost and is working on Driverless AI with H2O.ai.
JavaOne 2016: Code Generation with JavaCompiler for Fun, Speed and Business P...Juan Cruz Nores
On-the-fly bytecode generation is generally known to be super efficient, but also super difficult to implement and debug. Instead of trying to generate bytecode for the JVM, you can leverage the built-in Java compiler; generate Java code as a string, compile that to bytecode and then have that executed. This gives you better code efficiency, is easier to implement, and is straight-forward to debug. We’ll cover on-the-fly code generation, execution and debugging, working with HotSpot and G1 using dynamic code, as well as how to optimize for engineer implementation time; maximum gain in minimum time. We’ll use practical examples and code snippets, so you can be ready to make the core processing for your business 10x faster.
Learning Predictive Modeling with TSA and KaggleYvonne K. Matos
Ever wanted to do a challenging data science project but feel like you don’t have enough experience? Just go for it! Diving into a 3 TB, 3D image dataset has been my best learning experience. I want to share this deep learning project with you, and tips for overcoming challenges along the way.
Feature Engineering in H2O Driverless AI - Dmitry Larko - H2O AI World London...Sri Ambati
This talk was recorded in London on October 30th, 2018 and can be viewed here: https://youtu.be/d6UMEmeXB6o
In his talk Dmitry is going to cover common feature engineering techniques used to build robust machine learning models as well as some not widely known/used approaches.
Bio: Senior Data Scientist at H2O.ai, Dmitry Larko also is a former #25 Kaggle Grandmaster and loves to use his machine learning and data science skills in Kaggle Competitions and predictive analytics software development. He has more than 15 years of experience in information technology. Post his masters in computer information systems from Krasnoyarsk State Technical University (KSTU), he started his career in data warehousing and business intelligence and gradually moved to big data and data science. He holds a lot of experience in predictive analytics in a wide array of domains and tasks. Prior to H2O.ai, Dmitry held the position of SAP BW Developer at Chevron, Data Scientist at EPAM, and that of Lead Software Engineer with the Russian Federation.
Feature Engineering for ML - Dmitry Larko, H2O.aiSri Ambati
This talk was given at H2O World 2018 NYC and can be viewed here: https://youtu.be/wcFdmQSX6hM
Description:
In this talk, Dmitry shares his approach to feature engineering which he used successfully in various Kaggle competitions. He covers common techniques used to convert your features into numeric representation used by ML algorithms.
Speaker's Bio:
Dmitry has more than 10 years of experience in IT. Starting with data warehousing and BI, now in big data and data science. He has a lot of experience in predictive analytics software development for different domains and tasks. He is also a Kaggle Grandmaster who loves to use his machine learning and data science skills on Kaggle competitions.
Leakage in Meta Modeling And Its Connection to HCC Target-Encoding - Mathias ...Sri Ambati
Presented at #H2OWorld 2017 in Mountain View, CA.
Enjoy the video: https://youtu.be/Q0QmziFcfU0.
Learn more about H2O.ai: https://www.h2o.ai/.
Follow @h2oai: https://twitter.com/h2oai.
- - -
Abstract:
This talk covers the problems of leakage in a standard k-fold meta modeling (ensembling / stacking) setup, the connection to HCC target encoding and how to properly deal with the induced consequences in order to detect and finally avoid overfitting.
Mathias' Bio:
A Kaggle Grandmaster and a Data Scientist at H2O.ai, Mathias Müller holds an AI and ML focused diploma (eq. M.Sc.) in computer science from Humboldt University in Berlin. During his studies, he keenly worked on computer vision in the context of bio-inspired visual navigation of autonomous flying quadrocopters. Prior to H2O.ai, he as a machine learning engineer for FSD Fahrzeugsystemdaten GmbH in the automotive sector. His stint with Kaggle was a chance encounter as he stumbled upon the data competition platform while looking for a more ML-focused platform as compared to TopCoder. This is where he entered his first predictive modeling competition and climbed up the ladder to be a Grandmaster. He is an active contributor to XGBoost and is working on Driverless AI with H2O.ai.
JavaOne 2016: Code Generation with JavaCompiler for Fun, Speed and Business P...Juan Cruz Nores
On-the-fly bytecode generation is generally known to be super efficient, but also super difficult to implement and debug. Instead of trying to generate bytecode for the JVM, you can leverage the built-in Java compiler; generate Java code as a string, compile that to bytecode and then have that executed. This gives you better code efficiency, is easier to implement, and is straight-forward to debug. We’ll cover on-the-fly code generation, execution and debugging, working with HotSpot and G1 using dynamic code, as well as how to optimize for engineer implementation time; maximum gain in minimum time. We’ll use practical examples and code snippets, so you can be ready to make the core processing for your business 10x faster.
Learning Predictive Modeling with TSA and KaggleYvonne K. Matos
Ever wanted to do a challenging data science project but feel like you don’t have enough experience? Just go for it! Diving into a 3 TB, 3D image dataset has been my best learning experience. I want to share this deep learning project with you, and tips for overcoming challenges along the way.
In this talk, Dmitry shares his approach to feature engineering which he used successfully in various Kaggle competitions. He covers common techniques used to convert your features into numeric representation used by ML algorithms.
Feature Engineering - Getting most out of data for predictive models - TDC 2017Gabriel Moreira
How should data be preprocessed for use in machine learning algorithms? How to identify the most predictive attributes of a dataset? What features can generate to improve the accuracy of a model?
Feature Engineering is the process of extracting and selecting, from raw data, features that can be used effectively in predictive models. As the quality of the features greatly influences the quality of the results, knowing the main techniques and pitfalls will help you to succeed in the use of machine learning in your projects.
In this talk, we will present methods and techniques that allow us to extract the maximum potential of the features of a dataset, increasing flexibility, simplicity and accuracy of the models. The analysis of the distribution of features and their correlations, the transformation of numeric attributes (such as scaling, normalization, log-based transformation, binning), categorical attributes (such as one-hot encoding, feature hashing, Temporal (date / time), and free-text attributes (text vectorization, topic modeling).
Python, Python, Scikit-learn, and Spark SQL examples will be presented and how to use domain knowledge and intuition to select and generate features relevant to predictive models.
Piotr Mirowski - Review Autoencoders (Deep Learning) - CIUUK14Daniel Lewis
Piotr Mirowski (of Microsoft Bing London) presented Review of Auto-Encoders to the Computational Intelligence Unconference 2014, with our Deep Learning stream. These are his slides. Original link here: https://piotrmirowski.files.wordpress.com/2014/08/piotrmirowski_ciunconf_2014_reviewautoencoders.pptx
He also has Matlab-based tutorial on auto-encoders available here:
https://github.com/piotrmirowski/Tutorial_AutoEncoders/
Feature Engineering - Getting most out of data for predictive modelsGabriel Moreira
How should data be preprocessed for use in machine learning algorithms? How to identify the most predictive attributes of a dataset? What features can generate to improve the accuracy of a model?
Feature Engineering is the process of extracting and selecting, from raw data, features that can be used effectively in predictive models. As the quality of the features greatly influences the quality of the results, knowing the main techniques and pitfalls will help you to succeed in the use of machine learning in your projects.
In this talk, we will present methods and techniques that allow us to extract the maximum potential of the features of a dataset, increasing flexibility, simplicity and accuracy of the models. The analysis of the distribution of features and their correlations, the transformation of numeric attributes (such as scaling, normalization, log-based transformation, binning), categorical attributes (such as one-hot encoding, feature hashing, Temporal (date / time), and free-text attributes (text vectorization, topic modeling).
Python, Python, Scikit-learn, and Spark SQL examples will be presented and how to use domain knowledge and intuition to select and generate features relevant to predictive models.
How to Build your Training Set for a Learning To Rank Project - HaystackSease
Presented by Alessandro Benedetti of Sease, Learning to Rank (LTR) is the application of machine learning techniques (typically supervised), in the formulation of ranking models for information retrieval systems.
With LTR becoming more and more popular, organizations struggle with the problem of how to collect and structure relevance signals necessary to train their ranking models.
This talk is a technical guide to explore and master various techniques to generate your training set(s) correctly and efficiently.
Expect to learn how to :
- model and collect the necessary feedback from the users (implicit or explicit)
- calculate for each training sample a relevance label that is meaningful and not ambiguous (Click Through Rate, Sales Rate ...)
- transform the raw data collected in an effective training set (in the numerical vector format most of the LTR training libraries expect)
Join us as we explore real-world scenarios and dos and don'ts from the e-commerce industry.
Neural Networks in the Wild: Handwriting RecognitionJohn Liu
Demonstration of linear and neural network classification methods for the problem of offline handwriting recognition using the NIST SD19 Dataset. Tutorial on building neural networks in Pylearn2 without YAML. iPython notebook located at nbviewer.ipython.org/github/guard0g/HandwritingRecognition/tree/master/Handwriting%20Recognition%20Workbook.ipynb
SPARKNaCl: A verified, fast cryptographic libraryAdaCore
SPARKNaCl https://github.com/rod-chapman/SPARKNaCl is a new, freely-available, verified and fast reference implementation of the NaCl cryptographic API, based on the TweetNaCl distribution. It has a fully automated, complete and sound proof of type-safety and several key correctness properties. In addition, the code is surprisingly fast - out-performing TweetNaCl's C implementation on an Ed25519 Sign operation by a factor of 3 at all optimisation levels on a 32-bit RISC-V bare-metal machine. This talk will concentrate on how "Proof Driven Optimisation" can result in code that is both correct and fast.
Anil Thomas dives deep into the field of Deep Learning and focuses on object recognition. This talk will start with a general overview of how to use neon, Convolutional Neural Networks (CNN) and applying neon to an object recognition Kaggle problem. The talk is followed by a workshop highlighting neon, an open source python based deep learning framework that has been built from the ground up for speed and ease of use.
In this talk, Dmitry shares his approach to feature engineering which he used successfully in various Kaggle competitions. He covers common techniques used to convert your features into numeric representation used by ML algorithms.
Feature Engineering - Getting most out of data for predictive models - TDC 2017Gabriel Moreira
How should data be preprocessed for use in machine learning algorithms? How to identify the most predictive attributes of a dataset? What features can generate to improve the accuracy of a model?
Feature Engineering is the process of extracting and selecting, from raw data, features that can be used effectively in predictive models. As the quality of the features greatly influences the quality of the results, knowing the main techniques and pitfalls will help you to succeed in the use of machine learning in your projects.
In this talk, we will present methods and techniques that allow us to extract the maximum potential of the features of a dataset, increasing flexibility, simplicity and accuracy of the models. The analysis of the distribution of features and their correlations, the transformation of numeric attributes (such as scaling, normalization, log-based transformation, binning), categorical attributes (such as one-hot encoding, feature hashing, Temporal (date / time), and free-text attributes (text vectorization, topic modeling).
Python, Python, Scikit-learn, and Spark SQL examples will be presented and how to use domain knowledge and intuition to select and generate features relevant to predictive models.
Piotr Mirowski - Review Autoencoders (Deep Learning) - CIUUK14Daniel Lewis
Piotr Mirowski (of Microsoft Bing London) presented Review of Auto-Encoders to the Computational Intelligence Unconference 2014, with our Deep Learning stream. These are his slides. Original link here: https://piotrmirowski.files.wordpress.com/2014/08/piotrmirowski_ciunconf_2014_reviewautoencoders.pptx
He also has Matlab-based tutorial on auto-encoders available here:
https://github.com/piotrmirowski/Tutorial_AutoEncoders/
Feature Engineering - Getting most out of data for predictive modelsGabriel Moreira
How should data be preprocessed for use in machine learning algorithms? How to identify the most predictive attributes of a dataset? What features can generate to improve the accuracy of a model?
Feature Engineering is the process of extracting and selecting, from raw data, features that can be used effectively in predictive models. As the quality of the features greatly influences the quality of the results, knowing the main techniques and pitfalls will help you to succeed in the use of machine learning in your projects.
In this talk, we will present methods and techniques that allow us to extract the maximum potential of the features of a dataset, increasing flexibility, simplicity and accuracy of the models. The analysis of the distribution of features and their correlations, the transformation of numeric attributes (such as scaling, normalization, log-based transformation, binning), categorical attributes (such as one-hot encoding, feature hashing, Temporal (date / time), and free-text attributes (text vectorization, topic modeling).
Python, Python, Scikit-learn, and Spark SQL examples will be presented and how to use domain knowledge and intuition to select and generate features relevant to predictive models.
How to Build your Training Set for a Learning To Rank Project - HaystackSease
Presented by Alessandro Benedetti of Sease, Learning to Rank (LTR) is the application of machine learning techniques (typically supervised), in the formulation of ranking models for information retrieval systems.
With LTR becoming more and more popular, organizations struggle with the problem of how to collect and structure relevance signals necessary to train their ranking models.
This talk is a technical guide to explore and master various techniques to generate your training set(s) correctly and efficiently.
Expect to learn how to :
- model and collect the necessary feedback from the users (implicit or explicit)
- calculate for each training sample a relevance label that is meaningful and not ambiguous (Click Through Rate, Sales Rate ...)
- transform the raw data collected in an effective training set (in the numerical vector format most of the LTR training libraries expect)
Join us as we explore real-world scenarios and dos and don'ts from the e-commerce industry.
Neural Networks in the Wild: Handwriting RecognitionJohn Liu
Demonstration of linear and neural network classification methods for the problem of offline handwriting recognition using the NIST SD19 Dataset. Tutorial on building neural networks in Pylearn2 without YAML. iPython notebook located at nbviewer.ipython.org/github/guard0g/HandwritingRecognition/tree/master/Handwriting%20Recognition%20Workbook.ipynb
SPARKNaCl: A verified, fast cryptographic libraryAdaCore
SPARKNaCl https://github.com/rod-chapman/SPARKNaCl is a new, freely-available, verified and fast reference implementation of the NaCl cryptographic API, based on the TweetNaCl distribution. It has a fully automated, complete and sound proof of type-safety and several key correctness properties. In addition, the code is surprisingly fast - out-performing TweetNaCl's C implementation on an Ed25519 Sign operation by a factor of 3 at all optimisation levels on a 32-bit RISC-V bare-metal machine. This talk will concentrate on how "Proof Driven Optimisation" can result in code that is both correct and fast.
Anil Thomas dives deep into the field of Deep Learning and focuses on object recognition. This talk will start with a general overview of how to use neon, Convolutional Neural Networks (CNN) and applying neon to an object recognition Kaggle problem. The talk is followed by a workshop highlighting neon, an open source python based deep learning framework that has been built from the ground up for speed and ease of use.
Building LLM Solutions using Open Source and Closed Source Solutions in Coher...Sri Ambati
Sandeep Singh, Head of Applied AI Computer Vision, Beans.ai
H2O Open Source GenAI World SF 2023
In the modern era of machine learning, leveraging both open-source and closed-source solutions has become paramount for achieving cutting-edge results. This talk delves into the intricacies of seamlessly integrating open-source Large Language Model (LLM) solutions like Vicuna, Falcon, and Llama with industry giants such as ChatGPT and Google's Palm. As the demand for fine-tuned and specialized datasets grows, it is imperative to understand the synergy between these tools. Attendees will gain insights into best practices for building and enriching datasets tailored for fine-tuning tasks, ensuring that their LLM projects are both robust and efficient. Through real-world examples and hands-on demonstrations, this talk will equip attendees with the knowledge to harness the power of both open and closed-source tools in a coherent and effective manner.
Patrick Hall, Professor, AI Risk Management, The George Washington University
H2O Open Source GenAI World SF 2023
Language models are incredible engineering breakthroughs but require auditing and risk management before productization. These systems raise concerns about toxicity, transparency and reproducibility, intellectual property licensing and ownership, disinformation and misinformation, supply chains, and more. How can your organization leverage these new tools without taking on undue or unknown risks? While language models and associated risk management are in their infancy, a small number of best practices in governance and risk are starting to emerge. If you have a language model use case in mind, want to understand your risks, and do something about them, this presentation is for you!
Dr. Alexy Khrabrov, Open Source Science Community Director, IBM
H2O Open Source GenAI World SF 2023
In this talk, Dr. Alexy Khrabrov, recently elected Chair of the new Generative AI Commons at Linux Foundation for AI & Data, outlines the OSS AI landscape, challenges, and opportunities. With new models and frameworks being unveiled weekly, one thing remains constant: community building and validation of all aspects of AI is key to reliable and responsible AI we can use for business and society needs. Industrial AI is one key area where such community validation can prove invaluable.
Michelle Tanco, Head of Product, H2O.ai
H2O Open Source GenAI World SF 2023
Learn how the makers at H2O.ai are building internal tools to solve real use cases using H2O Wave and h2oGPT. We will walk through an end-to-end use case and discuss how to incorporate business rules and generated content to rapidly develop custom AI apps using only Python APIs.
Applied Gen AI for the Finance Vertical Sri Ambati
Megan Kurka, Vice President, Customer Data Scientist, H2O.ai
H2O Open Source GenAI World SF 2023
Discover the transformative power of Applied Gen AI. Learn how the H2O team builds customized applications and workflows that integrate capabilities of Gen AI and AutoML specifically designed to address and enhance financial use cases. Explore real world examples, learn best practices, and witness firsthand how our innovative solutions are reshaping the landscape of finance technology.
Practitioner's Guide to LLMs: Exploring Use Cases and a Glimpse Beyond Curren...Sri Ambati
Pascal Pfeiffer, Principal Data Scientist, H2O.ai
H2O Open Source GenAI World SF 2023
This talk dives into the expansive ecosystem of Large Language Models (LLMs), offering practitioners an insightful guide to various relevant applications, from natural language understanding to creative content generation. While exploring use cases across different industries, it also honestly addresses the current limitations of LLMs and anticipates future advancements.
Introducción al Aprendizaje Automatico con H2O-3 (1)Sri Ambati
En esta reunión virtual, damos una introducción a la plataforma de aprendizaje automático de código abierto número 1, H2O-3 y te mostramos cómo puedes usarla para desarrollar modelos para resolver diferentes casos de uso.
From Rapid Prototypes to an end-to-end Model Deployment: an AI Hedge Fund Use...Sri Ambati
Numerai is an open, crowd-sourced hedge fund powered by predictions from data scientists around the world. In return, participants are rewarded with weekly payouts in crypto.
In this talk, Joe will give an overview of the Numerai tournament based on his own experience. He will then explain how he automates the time-consuming tasks such as testing different modelling strategies, scoring new datasets, submitting predictions to Numerai as well as monitoring model performance with H2O Driverless AI and R.
AI Foundations Course Module 1 - Shifting to the Next Step in Your AI Transfo...Sri Ambati
In this session, you will learn about what you should do after you’ve taken an AI transformation baseline. Over the span of this session, we will discuss the next steps in moving toward AI readiness through alignment of talent and tools to drive successful adoption and continuous use within an organization.
To find additional videos on AI courses, earn badges, join the courses at H2O.ai Learning Center: https://training.h2o.ai/products/ai-foundations-course
To find the Youtube video about this presentation: https://youtu.be/K1Cl3x3rd8g
Speaker:
Chemere Davis (H2O.ai - Senior Data Scientist Training Specialist)
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
Techniques to optimize the pagerank algorithm usually fall in two categories. One is to try reducing the work per iteration, and the other is to try reducing the number of iterations. These goals are often at odds with one another. Skipping computation on vertices which have already converged has the potential to save iteration time. Skipping in-identical vertices, with the same in-links, helps reduce duplicate computations and thus could help reduce iteration time. Road networks often have chains which can be short-circuited before pagerank computation to improve performance. Final ranks of chain nodes can be easily calculated. This could reduce both the iteration time, and the number of iterations. If a graph has no dangling nodes, pagerank of each strongly connected component can be computed in topological order. This could help reduce the iteration time, no. of iterations, and also enable multi-iteration concurrency in pagerank computation. The combination of all of the above methods is the STICD algorithm. [sticd] For dynamic graphs, unchanged components whose ranks are unaffected can be skipped altogether.
Levelwise PageRank with Loop-Based Dead End Handling Strategy : SHORT REPORT ...Subhajit Sahu
Abstract — Levelwise PageRank is an alternative method of PageRank computation which decomposes the input graph into a directed acyclic block-graph of strongly connected components, and processes them in topological order, one level at a time. This enables calculation for ranks in a distributed fashion without per-iteration communication, unlike the standard method where all vertices are processed in each iteration. It however comes with a precondition of the absence of dead ends in the input graph. Here, the native non-distributed performance of Levelwise PageRank was compared against Monolithic PageRank on a CPU as well as a GPU. To ensure a fair comparison, Monolithic PageRank was also performed on a graph where vertices were split by components. Results indicate that Levelwise PageRank is about as fast as Monolithic PageRank on the CPU, but quite a bit slower on the GPU. Slowdown on the GPU is likely caused by a large submission of small workloads, and expected to be non-issue when the computation is performed on massive graphs.
Adjusting OpenMP PageRank : SHORT REPORT / NOTESSubhajit Sahu
For massive graphs that fit in RAM, but not in GPU memory, it is possible to take
advantage of a shared memory system with multiple CPUs, each with multiple cores, to
accelerate pagerank computation. If the NUMA architecture of the system is properly taken
into account with good vertex partitioning, the speedup can be significant. To take steps in
this direction, experiments are conducted to implement pagerank in OpenMP using two
different approaches, uniform and hybrid. The uniform approach runs all primitives required
for pagerank in OpenMP mode (with multiple threads). On the other hand, the hybrid
approach runs certain primitives in sequential mode (i.e., sumAt, multiply).
Adjusting primitives for graph : SHORT REPORT / NOTESSubhajit Sahu
Graph algorithms, like PageRank Compressed Sparse Row (CSR) is an adjacency-list based graph representation that is
Multiply with different modes (map)
1. Performance of sequential execution based vs OpenMP based vector multiply.
2. Comparing various launch configs for CUDA based vector multiply.
Sum with different storage types (reduce)
1. Performance of vector element sum using float vs bfloat16 as the storage type.
Sum with different modes (reduce)
1. Performance of sequential execution based vs OpenMP based vector element sum.
2. Performance of memcpy vs in-place based CUDA based vector element sum.
3. Comparing various launch configs for CUDA based vector element sum (memcpy).
4. Comparing various launch configs for CUDA based vector element sum (in-place).
Sum with in-place strategies of CUDA mode (reduce)
1. Comparing various launch configs for CUDA based vector element sum (in-place).
Enhanced Enterprise Intelligence with your personal AI Data Copilot.pdfGetInData
Recently we have observed the rise of open-source Large Language Models (LLMs) that are community-driven or developed by the AI market leaders, such as Meta (Llama3), Databricks (DBRX) and Snowflake (Arctic). On the other hand, there is a growth in interest in specialized, carefully fine-tuned yet relatively small models that can efficiently assist programmers in day-to-day tasks. Finally, Retrieval-Augmented Generation (RAG) architectures have gained a lot of traction as the preferred approach for LLMs context and prompt augmentation for building conversational SQL data copilots, code copilots and chatbots.
In this presentation, we will show how we built upon these three concepts a robust Data Copilot that can help to democratize access to company data assets and boost performance of everyone working with data platforms.
Why do we need yet another (open-source ) Copilot?
How can we build one?
Architecture and evaluation
2. 6x
"Coming up with features is difficult, time-consuming, requires expert knowledge. “Applied machine
learning” is basically feature engineering." ~ Andrew Ng
Feature Engineering
“… some machine learning projects succeed and some fail. What makes the difference? Easily the
most important factor is the features used.” ~ Pedro Domingos
“Good data preparation and feature engineering is integral to better prediction” ~ Marios Michailidis
(KazAnova), Kaggle GrandMaster, Kaggle #2, former #1
”you have to turn your inputs into things the algorithm can understand” ~ Shayne Miel, answer to
“What is the intuitive explanation of feature engineering in machine learning?”
13. Categorical Encoding – Target mean encoding 𝑓 𝑛 example
6x
x = frequency
k = inflection
point
f = steepness
https://www.researchgate.net/publication/220520258_A_Preprocessing_Scheme_for_High-Cardinality_Categorical_Attributes_in_Classification_and_Prediction_Problems
14. Categorical Encoding - Target mean encoding - Smoothing
6x
Feature Outcome
A 1
A 0
A 1
A 1
B 1
B 1
B 0
C 1
C 1
x level dataset 𝜆
A 4 0.75 0.77 0.99 0.99*0.75 + 0.01*0.77 = 0.7502
B 3 0.66 0.77 0.98 0.98*0.66 + 0.02*0.77 = 0.6622
C 2 1.00 0.77 0.5 0.5*1.0 + 0.5*0.77 = 0.885
𝜆 =
1
1 + exp (−
𝑥 − 2
0.25
)
x level dataset 𝜆
A 4 0.75 0.77 0.98 0.98*0.75 + 0.01*0.77 = 0.7427
B 3 0.66 0.77 0.5 0.5*0.66 + 0.5*0.77 = 0.715
C 2 1.00 0.77 0.017 0.017*1.0 + 0.983*0.77 = 0.773
𝜆 =
1
1 + exp (−
𝑥 − 3
0.25
)