Рубрика: AI

Image by Author

We’re 3 months into the new year. Wow, time is going by fast. With that being said, how many of you can say that you are on target to reach your goals, if it’s Q1 goals, learning goals, etc?

It’s hard. It’s hard to stay on top of everything and achieve every goal.

If learning data science was one of your 2024 goals, KDNuggets is here to help you so your journey to learning new skills and shifting careers is smooth.

I have put together a list of FREE online data science courses that will help you build a solid foundation of data science knowledge, skills and best practices to have a great career in the data world.

Understanding Data Science

From: DataCamp

Link: Understanding Data Science

If you’re new to the data science world, the first thing you want to get your head around is the basics. With no coding involved, this free course will define what data science is.

You will dive deeper into what a data science workflow is and how data science is applied to real-world problems. Once you have a good understanding of the field, you will end off will learning about different roles within the data science field.

Introduction to Programming with Python

From: Harvard University

Link: Introduction to Programming with Python

Have you decided to choose Python as your programming language choice? Great idea.

It’s been a popular programming language for a while now and now you can learn it with this self-paced course which will take you roughly 10 weeks to complete. This course has been specifically designed for students who have no prior experience or knowledge when it comes to programming and wish to transition into the data science world by learning Python.

The following topics are covered: functions, variables, conditionals, loops, exceptions, libraries, unit tests, file I/O, regular expressions, object-oriented programming, and more.

Python Data Science Toolbox

From: DataCamp

Link: Python Data Science Toolbox

As it’s been the most popular programming language for a few years now, there is no harm in perfecting the Python programming language.

With no coding experience or skill needed, Part 1 of the Python Data Science toolbox course teaches you how to effectively analyze and visualise your data. With 13 videos included in this course, you will go from data manipulation to plotting data with Matplotlib.

Data Science: R Basics

From: Harvard University

Link: Data Science: R Basics

Maybe you haven’t gone for Python and maybe you have decided to choose R as your programming language. Regardless of what you decide to go ahead with — it is always good to start with the basics. Harvard University offers a Data Science: R Basics course that helps you to build a solid foundation in the R programming language — from learning how to wrangle, analyze, and visualize data.

The course is free; however, you can pay for a verified certificate for $149.

Statistical Learning

From: Stanford Online

Link: Statistical Learning

I say it all the time and I will say it again — it is very important to learn about statistics in data science. This Statistical Learning course by edX will provide you with the main tools used in statistical modelling and data science.

It covers the following topics: an overview of statistical learning, linear regression, classification, resampling methods, linear model selection and regularization, moving beyond linearity, tree-based methods, support vector machines, deep learning, survival modelling, unsupervised learning, and multiple testing.

Data Analytics

From: Google

Link: Google Data Analytics

You have probably heard about this course a lot — it is very popular. It consists of 8 sections, where you will learn about the day-to-day use of data, best practices and processes that you should expect in your new data science/analytics jobs.

You will learn how to clean and organize data for the analysis process and make calculations using spreadsheets, SQL and R programming. It doesn’t stop there, you will further your analytical skills by creating data visualizations and also learning about tools such as Tableau.

Machine Learning Specialization

From: Coursera

Link: Machine Learning Specialization

This course has been put together by Andrew Ng, Founder & CEO of Landing AI, Founder of deeplearning.ai, and Co-Chairman and Co-Founder of Coursera. He has constructed a machine learning specialization made up of 3 courses:

1. Supervised Machine Learning: Regression and Classification
2. Advanced Learning Algorithms
3. Unsupervised Learning, Recommenders, Reinforcement Learning

These courses are free; however, there is a fee if you wish to get certified.

Wrapping it up

These 7 courses help you build skills in different aspects of data science, to ensure you have the skills and knowledge you require to be great at your data science job.

For example, working on being proficient in the Python programming language, or the R programming language. Understand the importance of learning statistics in data science, and how it affects the analytics process. Last but not least, put everything together and apply it to machine learning, for example, regression and recommenders.

Happy Learning!

Nisha Arya is a data scientist, freelance technical writer, and an editor and community manager for KDnuggets. She is particularly interested in providing data science career advice or tutorials and theory-based knowledge around data science. Nisha covers a wide range of topics and wishes to explore the different ways artificial intelligence can benefit the longevity of human life. A keen learner, Nisha seeks to broaden her tech knowledge and writing skills, while helping guide others.

More On This Topic

• 5 FREE Courses on AI with Microsoft for 2024
• Top Free Data Science Online Courses for 2023
• KDnuggets News, April 6: 8 Free MIT Courses to Learn Data Science…
• 8 Free MIT Courses to Learn Data Science Online
• Top 6 Data Science Online Courses in 2021
• 6 Best Free Online Courses to Jumpstart Your Learning of SQL

NetApp in collaboration with NVIDIA, today announced a groundbreaking advancement in generative AI applications, enabling secure and private data interaction across the hybrid multicloud. This partnership leverages NVIDIA NeMo Retriever microservices, a component of the NVIDIA AI Enterprise software platform, to unlock exabytes of data stored on NetApp’s intelligent data infrastructure.

Enterprises now have the capability to securely “talk to their data,” accessing proprietary business insights without compromising data security or privacy. This innovation addresses the challenges enterprises face in leveraging large language models for internal applications, such as chatbots and co-pilots, without risking data exposure.

NetApp and NVIDIA’s solution simplifies the process, allowing any data stored on NetApp ONTAP—whether on-premises or in the cloud—to be queried through simple prompts. This ensures that enterprises maintain control over their data while benefiting from AI-driven insights.

“As the leader in unstructured data management, NetApp makes data infrastructure intelligent to securely turbocharge AI innovation,” said George Kurian, CEO at NetApp. “AI is defining a new data-driven era and NetApp and NVIDIA are at the forefront of ensuring success for customers as they adopt and deploy AI.”

Jensen Huang, Founder and CEO at NVIDIA, emphasized the transformative potential of retrieval-augmented generation, stating, “Together, NVIDIA and NetApp can help enterprises build accurate, intelligent generative AI applications that let companies talk to their data — and foster a new wave of business opportunity.”

This collaboration not only reduces the friction, cost, and time to value for enterprises but also complements NetApp’s robust portfolio of AI offerings. Over 500 joint customers have already leveraged these solutions for AI model training and inference.

The “Talk to Your Data” capabilities were also showcased at NVIDIA GTC, highlighting the partnership’s commitment to driving AI innovation and data security for enterprises worldwide.

The post NetApp Unlocks Exabytes of Enterprise Data for GenAI with NVIDIA appeared first on Analytics India Magazine.

Equating NVIDIA chief Jensen Huang to ‘Don Corleone of AI’ may sound silly, but it’s probably not a stretch to do so (minus the mafia bit). Besides all the cool things unveiled at NVIDIA GTC, the biggest takeaway from Huang’s keynote was how indispensable NVIDIA is to all tech companies in the world.

Source: X

A Syndicate of AI Companies

Huang’s two-hour keynote speech not only disproved Moore’s law, but also showcased the various domains NVIDIA is impacting, including healthcare and humanoid robotics. The underlying theme of all these announcements was that every tech company needs NVIDIA.

The company’s strategic partnership with major big tech firms was further fortified at the conference. The display of the exhaustive list of NVIDIA partners could hardly fit on the screen, and you could probably play Bingo to identify every known player!

Partner list displayed on screen. Source: NVIDIA Keynote

With the announcement of the latest Blackwell GPU, NVIDIA also unveiled another list of all the companies lined up to use the same. “Blackwell will be the most successful product launch in our history, and I can’t wait to see that,” said Hunag, in his keynote.

Apart from the ‘big four’, which have been mentioned right at the top of the list, AI players such as Recursion, Cohere, Together AI, and others are there too. Elon Musk’s Tesla and X are also on the list. Interestingly, Indian data centre company Yotta, which recently received India’s first cluster of 4000 NVIDIA H100 GPUs, will receive Blackwell GPUs in October this year.

Companies that will use the newly unveiled Blackwell GPU. Source: NVIDIA Keynote

NVIDIA’s Four Horsemen

As highlighted on the top of the chart, NVIDIA’s big-four partners AWS, Google Cloud, Microsoft Azure, and Oracle Cloud had a separate mention.

Huang spoke about AWS gearing up for Blackwell, and how they are going to build the first GPU with secure AI. “NVIDIA and AWS engineers are also joining forces to co-develop an AI supercomputer, built exclusively on AWS for NVIDIA’s own AI R&D. It will feature 20K+ GB200 superchips, capable of processing a massive 414 exaflops,” said Adam Selipsky, the CEO of AWS.

Big-tech rivals, Google and Microsoft have ensured that they are not left behind.

Source: X

Google announced its adoption of the new NVIDIA Grace Blackwell AI computing platform, alongside the integration of the NVIDIA DGX Cloud service into Google Cloud. Furthermore, the NVIDIA H100-powered DGX Cloud platform is now universally accessible on Google Cloud.

“Together with NVIDIA, our team is committed to providing a highly accessible, open and comprehensive AI platform for ML developers,” said Thomas Kurian, the CEO of Google Cloud.

AI opportunist Microsoft, has been on a recent spree of expanding their AI efforts through strategic partnerships with emerging AI startups. The company even roped in Inflection AI’s founding team members, including Mustafa Suleyman to head Microsoft’s AI division, a position briefly given to Sam Altman during the OpenAI fiasco.

Microsoft Corp and NVIDIA enhanced their enduring partnership with new integrations, utilising the cutting-edge NVIDIA generative AI and Omniverse technologies across Microsoft Azure, Azure AI services, Microsoft Fabric, and Microsoft 365.

Tech giant Oracle, which is actively integrating generative AI capabilities into its OCI’s tech stack, has partnered with NVIDIA to provide accelerated computing and generative AI services. “Oracle is a great partner of ours for the NVIDIA DGX cloud and we’re also working together to accelerate something that’s really important to a lot of Oracle database companies,” said Huang.

Enabling enterprise solutions, NVIDIA even mentioned their growing support to SAP labs.

While big-tech service providers are taken care of, product companies are not left behind. Huang announced the availability of Omniverse on Apple Vision Pro headsets. Further, he even showed NVIDIA’s prowess with almost all major robotics companies for providing infra for their humanoid robots. Interestingly, the Tesla humanoid was not part of it.

Next Phase?

When the world is built on NVIDIA GPUs, symbiosis partnerships come as no surprise. Last month, the company hit a $2 trillion market value, and the growth is predicted to be on an upward trajectory. However, what’s interesting to note is that the pace at which NVIDIA is rising serves as an indicator of how other big tech companies, the so-called partners now, will also rise.

The CEO of Abacus AI, Bindu Reddy, has different levels of predictions for NVIDIA. Reddy believes that NVIDIA will likely diversify to other businesses. With the rise of alternatives and with ‘AMD becoming usable’, GPU prices will drop. “As with everything, compute will become a commodity where margins tend to be zero,” said Reddy.

With Amazon’s silicon innovation already on the way with AI chips, it is possible that the reliance on the tech giant can be reduced in the future. Google and Microsoft are also not far behind with their AI chip venture.

In the future, these big tech companies can rely on their chips for self-sustenance. This could change the course for NVIDIA, and partnership strategies can vary. Until then, NVIDIA will see all other companies allying with them. After all, you have to play ball with Don Corleone if you want to stay in the game.”

The post Why Tech CEOs are Kissing Jensen Huang’s Ring appeared first on Analytics India Magazine.

Jeff Dean, chief scientist at Google DeepMind, took to X to share that the company is starting to roll out API support for Gemini 1.5 Pro.

Gemini 1.5 Pro comes with a standard 128,000 token context window. It can process vast amounts of information in one go, including one hour of video, 11 hours of audio, codebases with over 30,000 lines of code, or over 700,000 words. In their research, Google also successfully tested up to 10 million tokens.

“We’ll be onboarding people to the API slowly at first, then ramping it up. In the meantime, developers can try out Gemini 1.5 Pro in the AI Studio UI right now,” said Dean.

Gemini 1.5, using Transformer and MoE architecture, combines the strengths of both models. Traditional Transformers function as one large network, whereas MoE divides models into smaller “expert” networks. Gemini 1.5 Pro excels in various tasks, such as analysing historical transcripts like Apollo 11’s mission and understanding silent movies. It efficiently processes extensive code, showcasing adaptability.

Notably, the Needle In A Haystack (NIAH) evaluation achieves a 99% success rate in locating specific facts within long texts. Its ability to learn in context, shown in the Machine Translation from One Book (MTOB) benchmark, establishes it as a leader in adaptive learning.

This development follows Google’s release of Gemini Ultra. Additionally, Google integrated generative AI features into Chrome and introduced the “Help Me Write” feature across all websites.

The post Google Introduces API Support for Gemini 1.5 Pro appeared first on Analytics India Magazine.

Large language models (LLMs) have revolutionized natural language processing (NLP) by excellently creating and understanding human-like text. However, these models often need to improve when it comes to basic arithmetic tasks. Despite their expertise in language, LLMs frequently require assistance with simple math calculations. This gap between language proficiency and mathematical skills has prompted researchers to investigate specialized models for arithmetic tasks.

In the fields of artificial intelligence and education, GOAT, which stands for Good at Arithmetic Tasks, has emerged as a remarkable development. Unlike traditional models, GOAT excels not only in NLP but also in solving complex mathematical problems. Imagine a model that effortlessly crafts expressive sentences while accurately solving complex equations. GOAT represents this unique combination, a skilled linguist and mathematician seamlessly integrated.

GOAT is a revolutionary AI model that excels at linguistic and numerical tasks. Unlike traditional language models, which focus mainly on generating and understanding text, GOAT outperforms them by demonstrating advanced mathematical problem-solving abilities. Its transition between these two domains marks a significant breakthrough in AI, opening opportunities for innovative applications in education, problem-solving, and other fields.

The GOAT Model

The GOAT model represents a significant advancement in artificial intelligence, specifically addressing the intersection of language understanding and mathematical reasoning. At its core, GOAT is a fine-tuned LLaMA model, a specialized variant of LLMs designed explicitly for arithmetic tasks. Unlike generic LLMs, which excel in NLP but struggle with basic arithmetic, GOAT has undergone targeted fine-tuning to enhance its mathematical capabilities.

GOAT’s superiority lies in its ability to tackle a wide range of arithmetic tasks with high accuracy. Compared to the widely acclaimed GPT-4, GOAT consistently delivers superior results in addition, subtraction, multiplication, and division. Its fine-tuned architecture enables it to effectively handle numerical expressions, word problems, and mathematical reasoning. Whether calculating large numbers or solving complex equations, GOAT demonstrates a level of precision that sets it apart from its predecessors.

To achieve this skill, GOAT uses a synthetically generated dataset. This dataset comprises diverse arithmetic examples covering various difficulty levels, number ranges, and problem types. By training on this carefully curated data, GOAT learns to generalize across different scenarios, making it adept at handling real-world arithmetic challenges.

GOAT’s capabilities extend beyond simple addition and subtraction. It conquers complex arithmetic challenges across various domains. Whether algebraic expressions, word problems, or multi-step calculations, GOAT consistently outperforms its competitors. Its accuracy and efficiency set a new standard.

The PaLM-540B, a powerful language model, encounters tough competition from the GOAT. In direct comparisons, GOAT shows better accuracy and strength. It handles complex numbers expertly, surpassing other models. GOAT’s strength comes from its supervised fine-tuning. Even when dealing with very large numbers that would challenge most, GOAT performs significantly well. It performs addition and subtraction accurately, demonstrating its mathematical brilliance.

Tokenization of Numbers in GOAT: Enhancing Arithmetic Precision

GOAT demonstrates a remarkable ability to handle numerical tokens consistently. Tokenization breaks down input text into smaller units or tokens. In GOAT’s case, these tokens represent both words and numerical values. GOAT ensures uniform treatment of numbers—integers, decimals, or scientific notation. Each numeric token receives equal attention, regardless of context.

In addition, GOAT ensures precision in parsing numerical expressions. When GOAT encounters an arithmetic expression, it dissects it into tokens. For instance, the expression “2.14 + 2.618” becomes the sequence of tokens: [“2.14”, “+”, “2.618”].

GOAT’s understanding of numerical tokens enables accurate operations. It recognizes that “2.14” is a decimal, “+” is an addition operator, and “2.618” is another decimal. This consistent handling ensures GOAT does not confuse numerical values with linguistic elements.

Solving Word Problems with Precision

In word problems, GOAT’s tokenization plays a crucial role.

Consider: “If Alice has 6 apples and Bob gives her 4 more, how many apples does Alice have?”

GOAT identifies numeric tokens (“6” and “4”) and the relevant operation (“gives her”). It computes the result accurately: 6 + 4 = 10. Thus, by treating numbers as distinct tokens, GOAT avoids ambiguity.

Likewise, GOAT accurately handles large numbers and scientific notation by preserving high precision. GOAT’s tokenization extends to large numbers, such as “1,000,000” or “1.23e6” (scientific notation for 1.23 × 10^6). Whether parsing a million or dealing with exponents, GOAT maintains precision.

Training, Fine-tuning, and Open Source Availability

The GOAT model is trained using a supervised approach, learning from labeled data and explicit instructions. A crucial step in its training process involves fine-tuning, where a pre-trained model, such as a language model, is adapted to a specific task by updating its weights based on task-specific data.

GOAT employs guided instructions during fine-tuning, ensuring targeted guidance throughout the adaptation process and enabling the model to generalize effectively to out-of-distribution examples. LoRA, as part of this paradigm, facilitates Low-Rank Adaptation, which enhances the robustness of the model. By incorporating LoRA, GOAT effectively handles label noise and improves the quality of training data, enabling it to learn effectively from noisy or imperfectly labeled data.

In addition, the GOAT model and its pre-trained weights are available as open-source software. Researchers can access the GOAT repository containing the model architecture, training code, evaluation scripts, and the dataset used for its training. This open-source approach encourages collaboration, innovation, and exploration within the scientific community, facilitating advancements in natural language understanding.

Challenges and Possible Solutions

Due to its complexity, the GOAT model needs help handling large-number multiplication and division. To overcome this, GOAT employs several strategies. First, it decomposes complex operations into smaller steps, such as multiplying individual digits or estimating quotients.

Additionally, it classifies tasks based on learnability—basic arithmetic is directly fine-tuned, while complex tasks are broken down. Guided fine-tuning provides explicit instructions during training, and attention mechanisms enhance performance. Sequential learning and transfer from more straightforward tasks empower GOAT to tackle complex arithmetic problems effectively.

The Bottom Line

In conclusion, GOAT is a significant advancement in AI, combining language understanding and mathematical reasoning. Its exceptional ability to handle arithmetic tasks, fine-tuned approach, and attention to numerical tokens demonstrates incomparable versatility and precision. With its open-source availability and ongoing advancements, GOAT paves the way for innovative applications in education and problem-solving, promising a future of enhanced AI capabilities.