Рубрика: AI

Former Twitter engineers are building Particle, an AI-powered news reader, backed by $4.4M Sarah Perez @sarahintampa / 9 hours

A team led by former Twitter engineers is rethinking how AI can be used to help people process news and information. Particle.news, which entered into private beta over the weekend, is a new startup offering a personalized, “multi-perspective” news reading experience that not only leverages AI to summarize the news, but also aims to do so in a way that fairly compensates authors and publishers — or so is the claim.

While Particle hasn’t yet shared its business model, it arrives at a time when there’s a growing concern about the impact of AI on a rapidly shrinking news ecosystem. News that is summarized by AI could limit clicks to publishers’ websites, which means their ability to monetize via advertising would also be reduced.

The startup was founded last year by former Senior Director of Product Management at Twitter, Sara Beykpour, who worked on products like Twitter Blue, Twitter Video, and conversations, and who spearheaded the experimental app, twttr. She had been at Twitter from 2015 through 2021, growing her position from software engineering to that of a senior director of product management. Her co-founder is a former senior engineer at both Twitter and Tesla, Marcel Molina.

The premise behind Particle, as Beykpour explained last month, is to make it easier to keep up with news using AI.

“Sometimes it feels like headlines are all we have time for. We also want to understand more, but faster,” she wrote in an introduction to the startup on Threads. “We’re in the early stages of using AI to transform the way we interact with news.”

Using Particle, news readers are offered a quick, bulleted summary of the story, with information pulled from a variety of sources. However, when announcing the private beta, Beykpour noted that readers can either use the summary to get up to speed or can choose to go deeper to “learn about how a story has unfolded over time.”

The venture-backed startup has raised a total of $4.4 million in seed funding from Kindred Ventures and Adverb Ventures, as well as various angel investors, including Twitter and Medium co-founder Ev Williams and Behance founder, Scott Belsky. The round closed in April 2023.

Remarked Belsky on X, “Particle has become a daily app for me. It synthesizes the many articles (and angles) on any news topic, surfaces the key points as objectively as possible, and lets you dig further across many dimensions. In the era of abstraction ahead, great example of daily AI,” he wrote.

Particle offers a demo of its technology for logged-out users via its website, where articles are featured along with their summary, timestamp as to when they were last updated, and, in a small section at the bottom, the sources they draw from.

These sources pull from across the political spectrum and include big-name publishers like The New York Times, CNBC, the AP, ABC, CNN, Breitbart, The Guardian, The Washington Post, Politico, Fox News, USA Today, The Daily Caller, New York Post, The Hill, and others. International outlets are also pulled from, when relevant, the demos indicate. However, each bullet point is not linked to its original source or sources, which makes it difficult to fact-check the accuracy of the AI summary without delving into all the articles. (Key terms are, however, linked.) We noted, too, that the photograph accompanying a news summary is watermarked with the publisher’s logo.

Image Credits: Particle

The end product will likely differ, given that Particle is just now launching its private beta for testing and intends to offer a mobile app in the future, as it’s hiring for a senior iOS engineer.

A similar model of leveraging a variety of news sources and then employing AI to summarize, was recently employed by Artifact, the now-shuttered startup from Instagram’s co-founders. In its case, Artifact’s team curated the news sources upfront based on factors related to their integrity and quality. For example, the outlet had to be quick to make corrections, when wrong, and be transparent about their funding. We’re hoping to talk in more detail about how Particle vets its sources closer to a public launch.

Another AI-powered news app, Bulletin, also recently launched to tackle clickbait along with offering news summaries.

Given the interest in this space, what could make Particle stand out is its founding team. Arriving from Twitter, the co-founders have experienced what a real-time news ecosystem feels like, and have the technical and product experience to build a quality product. Whether or not publishers who feel that AI is eating into their space will feel “fairly compensated,” however, remains to be seen.

Adverb Ventures co-founder and managing director April Underwood praised Particle in a post on LinkedIn about the firm’s investment:

“We got the chance to back them just as we were completing our very first close for Fund 1 — we had to wait for our first capital call to hit to wire them the money!” she said on Sunday, adding that Adverb closed its $75 million Fund I just a couple of months ago. “Sara and Marcel are the kind of founders we dreamed of backing when we set out to build a new early-stage firm. They are going after a big problem space. They’ve got the skills to tackle big problems at a high level of product quality. And they can attract other talented folks to join them, and together invent a future consumers don’t know to ask for (yet),” Underwood wrote.

In an email with TechCrunch, Underwood explained the opportunity ahead:

In terms of the space, we believe AI is going to touch every aspect of people’s digital lives at work and at home. Couple that with the pre-existing conditions at play here — it’s hard to find breaking news from sources you can trust, and the social media landscape is rapidly evolving — and you have to believe that the way people consume news is going to be different a few years from now. Sara and Marcel are uniquely qualified to help people get the news they need in a modern way.

Beykpour tells TechCrunch that the idea for Particle came about because there are a lot of challenges around how people get news and stay up to date with what’s happening. However, the company is still talking to publishers about what they would need to feel fairly compensated by a model where AI is summarizing their work.

“Honestly, we’re figuring that out. We’re talking to and working with publishers now to figure out what the right model is,” she says. “But my goal here is to make it fair.”

The company expects to have more news to share on that front in the months ahead. In the meantime, Particle’s beta sign-up form is here.

Originally published 2/26/24. Updated after publication with additional comment from Underwood. 2/26/24, 4:30 p.m. ET. Updated, 2/29/24, 8:30 a.m. PT with more funding details.

Today, AWS’ involvement in the telecommunication industry is way beyond helping telcos migrate to the cloud. In the current age, telcos are rethinking their technology stacks to drive agility and are transitioning from being called telcos to techcos.

With an increasing demand for high-speed connectivity, seamless communication, and a myriad of services, telcos are exploring and leveraging the best of emerging technologies to not only meet these demands but to exceed customer expectations.

As a result, the telecommunications industry has surfaced as a pivotal sector in AWS’s strategic landscape. AWS is actively contributing to the innovation endeavours of telcos through prototyping, co-development, the creation of new business lines, supporting their go-to-market efforts and achieving their sustainability goals.

“We are actively participating in the transformation of these telecommunications companies, tailoring our support to their specific needs across a spectrum of infrastructure, including enterprise IT, data storage, network components, and value-added services.

“Simultaneously, through our partner community, we ensure the exposure of various network capabilities of telcos through APIs, addressing a key focus area for many telecom companies in the current landscape,” Jayanth Nagarajan, head, telecommunications industry – Asia Pacific & Japan, AWS, told AIM in an exclusive interaction.

Bringing GenAI to telcos

“While a global initiative is underway to develop LLMs for the telecommunications industry, led by Korea-based SK Telecom, AWS research shows that 65% of telecommunications companies anticipate using off-the-shelf [LLMs] by fine-tuning them with their own enterprise data,” he said.

With Bedrock, AWS presents a range of LLMs like Llama by Meta, Claude by Anthropic, and additional models sourced from ElevenLabs, Stability AI, and open-source models featured on Hugging Face, encompassing their proprietary Titan models.

According to Nagarajan, the hyperscaler is helping telcos across the globe leverage these models across three themes – networks, consumers, and employees. When it comes to network, AWS is delivering generative AI solutions across three areas – network observability, prediction and troubleshooting.

“Generative AI helps in network root cause analysis. In the case of an unfortunate event, the generative AI-powered solutions that we have, very quickly allow the network operators to identify the root cause as well as change things around,” he pointed out.

Generative AI is also being used by telcos for network optimisation. For instance, LLMs are automating the configuration of network elements such as routers, switches, and other devices in a telecommunications network.

“We’re also seeing applications on forecasting customer needs, suggesting retention offers and personalising recommendations via a customer-connected journey,” Nagarajan said.

Moreover, AWS generative AI solutions also play a crucial role in automating various tasks which include streamlining Request for Proposal (RFP) responses, offering real-time support for field technicians seeking prompt assistance without relying on call centres and automating processes such as code generation, debugging, and testing.

How Telcos are leveraging CodeWhisperer and AI Chips

Over the years, AWS has also released custom silicon chips optimised for the cloud. At AWS re:Invent 2023, the hyperscaler announced Trainium2, which delivers up to 4x faster training than its first generation, with deployment capability in EC2 UltraClusters of up to 100,000 chips.

AWS silicon chips, including Trainium and Inferentia, are built to ensure that the training and inference are done at a very competitive price performance approach, according to Nagarajan.

Sharing the example of Deutsche Telekom, a Germany-based telco, he said AWS’ Inferentia 2 chips allowed them to increase the volume of queries answered by 8%.

“It improved the acceptable answer rate by 10% and reduced the fabricated responses or hallucinations by 25%. Overall it achieved its performance targets of improving better latency as well as higher throughput.”

Similarly, British Telecom (BT) recently announced that they are now deploying Amazon CodeWhisperer, which writes code in an integrated development environment with Amazon Q, a generative AI chatbot designed explicitly for businesses announced by AWS at re:Invent.

“The developers at BT have seen code suggestion acceptance rates of 37%. So there were 100,000 lines of code that developers did not need to write by hand,” Nagarajan revealed.

Deploying 5G with AWS

Globally, telcos have deployed or are in the process of deploying 5G services across the territories they operate in. AWS is not just assisting telcos in deploying 5G networks but also in formulating their monetisation strategies.

Nagaranjan emphasised that, unlike the 4G standards released in 2008, which were not inherently aligned with the cloud, the 2019 release of the 5G standard was explicitly designed to adhere to cloud operating principles.

“In 2019, we released AWS Outposts, which allowed telcos to run 5G network workloads on the cloud, not just in our hyperscale availability regions, but also on their premises. We also made available AWS wavelength, which is our multi-access edge compute engagement that we put at the telcos 5G network to help them monetise 5G,” he said.

Operators such as Dish Network in the US have made remarkable strides, expanding from zero to covering 20% of the continental US with AWS, according to Nagaranjan.

“They have openly acknowledged that this achievement would not have been possible without AWS, enabling them to deploy a 5G network within a record time of 14 months. Presently, they have surpassed the 70% mark in coverage,” he said.

Monetising 5G

Moreover, AWS is helping telcos with their monetising strategy when it comes to 5G. Since telcos understand connectivity well, according to Nagaranjan, one theme that is significantly gaining momentum is private networks.

Indian IT giant TCS and AWS are working on a broad range of initiatives to enable enterprises to unlock the power of private 5G, Nagaranjan revealed.

“TCS and AWS are working with enterprises to deliver actual business outcomes atop these private 5G networks. AWS is complementing the effort with our deep community of partners along with AI/ML and analytics services at the edge to enable these industry 4.0 solutions,” he concluded.

The post How AWS is Helping Telcos Become Techcos appeared first on Analytics India Magazine.

A recent addition to the DeepLearning.AI course catalogue is Meta’s “Engineering with Llama 2.” This course allows you to explore prompt engineering using the company’s Llama 2 models. The course is tailored for beginners, requiring just one hour. Led by instructor Amit Sangani, senior director of partner engineering, Meta, the course is currently available for free for a limited time.

You can sign up for the course here.

Participants will gain insights into the best practices associated with prompting Llama 2 models, focusing on practical applications. The curriculum encourages interaction with three key models: Meta Llama 2 Chat, Code Llama, and Llama Guard. These models serve distinct purposes, ranging from conversation and coding assistance to content moderation through Llama Guard.

The primary learning objectives of the course involve familiarizing oneself with the Llama 2 collection, adopting best practices in prompt engineering, and building applications using these models. Through a simple API call, participants will explore the diverse outputs of the Llama 2 models, gaining a nuanced understanding of their capabilities.

The course instructs participants on leveraging Llama 2 models as personal assistants, providing guidance for day-to-day tasks. Additionally, it delves into advanced prompt engineering techniques such as few-shot prompting for sentiment analysis and chain-of-thought prompting for logical problem-solving. Code Llama is introduced as a collaborative partner for pair programming, facilitating both learning and code improvement.

The course also throws light on the responsible use of LLMs by incorporating Llama Guard, which screens user prompts and model responses for potentially harmful content. Importantly, participants are informed that Llama 2 models and their weights are available for free download, including quantized versions for local machine deployment. The course also encourages involvement in an active open-source community that utilizes Llama 2 for building diverse applications.

The post Andrew Ng Unveils New Free Course on Llama 2 with Meta appeared first on Analytics India Magazine.

Tech Mahindra has teamed up with Indosat Ooredoo Hutchison to build ‘Garuda,’ a Large Language Model (LLM) to preserve Bahasa Indonesia, the official and national language of Indonesia and its dialects.

Garuda will be built on the principles of Tech Mahindra’s indigenous LLM ‘Project Indus‘, a foundational model designed to converse in a multitude of Indic languages and dialects.

The IT giant signed a Memorandum of Understanding (MoU) at Mobile World Congress (MWC) 2024.

As part of this partnership, Tech Mahindra will leverage its technology expertise to gather and curate data in the Indonesian language, which will be pre-trained and released as a conversational model for Indosat.

Garuda will be developed with 16 billion original Bahasa tokens, providing 1.2 billion parameters to shape the model’s understanding of the Bahasa language. These parameters will influence how the model processes input and formulates output.

A beta version of the Garuda model will be released for testing by Indosat and Bahasa Indonesia speakers. The model will be further improved using RLHF (Reinforcement Learning from Human Feedback) techniques to ensure its robustness for conversation. Additionally, any specialized use cases will be developed using the LIMA (Less is More for Alignment) method.

“The LLM market is expected to reach 40.8 billion USD by 2029. In this direction, the emergence of LLMs such as Garuda and Indus can enable people and enterprises to communicate online in their local dialects and languages, creating new opportunities in the digital world.

“We believe that the model will significantly promote Indonesia’s linguistic diversity and unlock new business opportunities for enterprises in the region,” said Harshvendra Soin, President – Asia Pacific and Japan Business, Tech Mahindra.

The post Tech Mahindra to Build LLM for Indonesia on Project Indus Principles appeared first on Analytics India Magazine.

Image by Editor

Machine Learning is a massive field that has changed the world, and I bet it will grow. If you have followed the current trend, you can see AI and machine learning have slowly but surely integrated into people’s everyday lives, not just a tool for creating a business competitive edge but giving real value to people's lives.

With how popular machine learning has become, people will try to enter this field. However, it’s hard to find a voice that could lead these enthusiasts to do something to enter the machine learning field.

Luckily, many podcasts specifically discuss machine learning and the careers surrounding them. What are these podcasts? Let’s get into it.

1. This Week in Machine Learning & AI

The popular Machine Learning podcast hosted by Sam Charrington has been running since 2016 and has published hundreds of episodes. This Week in Machine Learning & AI is one of the longest-running machine learning podcasts and is still going strong. The podcast launches a new episode every week with various machine learning and AI expert guests coming to discuss the latest topics in the field.

The podcast topic discussion is often about the latest breakthrough in machine learning, such as new algorithms or methodology. But it could extend to anything related, such as implication, business usage, and ethical consideration.

The topic discussed might seem pretty advanced for the enthusiasts, but it could provide insight and inspiration for the future career you want. You'll gain insights not found in textbooks or papers, ensuring you stay up-to-date with the latest findings.

2. The Machine Learning Podcast

The Machine Learning Podcast hosted by Tobias Macey is a machine learning focus podcast that run since 2022 and produce a monthly episode for the listener. This podcast brings various machine learning guests to discuss important topics in machine learning and what happens behind the scenes.

The topic could be the latest algorithms, techniques, tools, companies, or anything about machine learning. What is important is that each hourly episode wants the listener to get value and deliver impact from the machine learning to the business.

It’s a good podcast for enthusiasts wanting to know what happens in the business and why machine learning is important within companies. Don’t miss these podcasts if you want to have advantages in the machine learning field.

3. Practical AI

The Practical AI podcast hosted by Chris Benson and Daniel Whitenack provides a great addition to your machine learning listening list. The podcast has provided weekly episodes since 2018 to give listeners a more accessible way to understand machine learning and AI.

Many topics cover practical ways and high levels of machine learning so the listener can work outside the industry. However, it heavily focuses on current trending topics such as LLM, so it might not cover everything that happens in the machine learning industry.

The podcast is great for the enthusiast who wants to understand a high level of things and wants to catch up with what is happening in the world.

4. Super Data Science Podcast

The Super Data Science Podcast hosted by Jon Krohn is not necessarily focused only on machine learning, as the topic encompasses the whole data science field. However, many episodes discuss beginner-friendly machine learning introduction and the related niche to teach the listeners better about the field.

The podcast has run since 2016 and still produces weekly hour-episode podcasts dedicated to the data science field. For the machine learning enthusiast, there are more than 100 episodes for you to listen to and learn from, so you don’t need to worry about exhausting all the episodes shortly.

5. Machine Learning Street Talk

The Machine Learning Street Talk podcast, which Tim Scarfe runs, is a top podcast that discusses many recent topics in machine learning and AI more casually. Sometimes, the episode could cover an in-depth discussion of the field, but sometimes, it is a debate between host and guest.

The podcast makes its promotion less serious than the other podcasts, so it’s great for enthusiasts afraid to enter the field because of the technical terms. You could listen to hundreds of episodes, although the schedule is not fixed, and the length could vary between 1 hour to 3 hours and more.

Conclusion

These are the five podcasts for you machine learning enthusiasts. The podcast would bring you a fresh perspective on machine learning from the expert and give you insight into navigating the field. If you have any more podcast recommendations, please share them in the comments.

Cornellius Yudha Wijaya is a data science assistant manager and data writer. While working full-time at Allianz Indonesia, he loves to share Python and Data tips via social media and writing media.

More On This Topic

• 5 Machine Learning Skills Every Machine Learning Engineer Should…
• The 6 Python Machine Learning Tools Every Data Scientist Should Know About
• KDnuggets News, May 25: The 6 Python Machine Learning Tools Every…
• Every Engineer Should and Can Learn Machine Learning
• Best Instagram Accounts to Follow for Data Science, Machine Learning & AI
• Top 5 AI Podcasts You Can't Miss in 2024

The Mobile World Congress (MWC), an annual event organised by the GSMA, showcased the latest in mobile technology, including smartphones, services, and advancements in 5G and artificial intelligence. Held in Barcelona, Spain, it is the largest exhibition for the mobile industry, attracting global participants from the tech community.

This year, it hosted a plethora of devices with AI integration being the common thread. One standout example of AI’s application was the Honor Magic V2, where eye-tracking technology allows users to interact with their device in a hands-free manner.

This feature, along with other AI-driven innovations presented at MWC, underscores the industry’s shift towards creating more personalised and efficient user experiences.

Here is a list of top 10 gadgets showcased at MWC this year.

HMD Barbie Phone

HMD, initially celebrated for reviving Nokia phones at MWC 2017, struggled to compete with giants like Samsung and Apple, shifting its focus to budget Android and feature phones. At MWC 2024, it announced its first profitable year in 2023 and introduced a rebranding strategy, adopting ‘Human Mobile Devices’ as its new moniker.

Its 2024 device lineup featured a classic Nokia model, and a unique Barbie flip phone developed in collaboration with Mattel, targeting a summer release as a pink, digital detox tool. As for the non-Barbie phones, the company has plans for those, too, though no details are available at the moment.

Motorola Debuts Smart Connect

Motorola showed off its innovative Adaptive Display concept phone, a departure from traditional designs with its bendable structure, allowing it to be bent backward.

Motorola also introduced Smart Connect, a collaborative effort with Lenovo that builds upon the Ready for platform. This new feature allows for wireless connection between a Motorola phone and nearby displays, including Lenovo tablets and Windows laptops available through the Microsoft Store, enhancing productivity and inter-device usability.

OnePlus Watch 2

In the spotlight at MWC 2024 was the OnePlus Watch 2, which boasts a significant improvement over its predecessor. A standout feature of the OnePlus Watch 2 is its dual operating system capability, powered by two distinct processors. It operates on Google Wear OS with the Qualcomm Snapdragon W5 Gen 1 chipset for demanding tasks such as navigation, music playback, and app usage.

The OnePlus Watch 2, priced at $300, is currently available for preorder and will officially go on sale on March 4. OnePlus is also offering a promotional discount of $50 for those trading in any watch, including analog models, towards the purchase.

A Transparent Laptop from Lenovo

Lenovo introduced a concept at MWC 2024, known as Project Crystal, a transparent laptop. While it’s not slated for immediate release, the concept showcases a glimpse into the future of laptop design. The laptop’s Micro-LED transparent screen offers a futuristic look, allowing users to see through the device while still providing a bright display for normal app usage.

However, this transparency means that others can see the user’s screen, posing privacy concerns. Lenovo mentioned the potential for adjusting the screen’s transmissivity to create an opaque layer for privacy, though such features were not demonstrated.

Samsung Galaxy Ring

Samsung unveiled its latest wearable, the Galaxy ring, at the Mobile World Congress. This is the first time it was showcased to the public. This smart ring, designed to monitor health data and provide insights based on daily and nightly metrics, will expand Samsung’s wearable market. The Galaxy ring can monitor temperature, heart rate, respiratory rate, sleep movement, and time taken to fall asleep.

Interestingly, the Galaxy ring will also offer payment capabilities, distinguishing it from other smart rings that focus solely on health or fitness tracking. The ring is available in black, gold, and silver, and comes in nine sizes, accompanied by a sizing kit. Its price in India starts from ₹24,599 and is set to be released later this year.

Honor Magic V2

Honor displayed its new devices, the Magic 6 Pro and Magic V2 RSR smartphones, and the MagicBook Pro 16 laptop, heavily emphasising AI features that aren’t necessarily driven by actual artificial intelligence. A notable demonstration featured the eye-tracking technology on the Magic 6 Pro, which allows users to expand notifications by simply looking at them.

This feature, expected to be added via a software update, was highlighted through an unusual demo where the technology was used to control a car (an Alfa Romeo), with options like Engine Start and Stop, Forward, and Backward, showcasing the potential of eye-tracking for hands-free device interaction.

TCL’s NXTPaper 5G and Portable 5G Dongle

TCL introduced a new addition to its NXTPaper range, the TCL 50 XL NXTPaper 5G, featuring a 6.8-inch screen with a 120-Hz refresh rate designed to mimic paper. Despite its modest specs, its $229 price point is aimed at readers preferring to engage with digital content on their phones.

Additionally, TCL unveiled the NXTPAPER 14 Pro, equipped with the same eye-friendly technology in a larger 14-inch display, targeting productivity users with its MediaTek Dimensity 8020 processor, 12 GB of RAM, a 12,000-mAh battery, and 256 GB of storage.

ZTE 5G+AI Eyewear-free 3D Tablet

Nubia introduced its latest devices on the ZTE stage, despite emphasising its independence from ZTE. The highlights include the Nubia Flip, the brand’s first foldable phone, featuring a 6.9-inch 120-Hz display that folds to a compact size and sports a unique circular screen on the front. Priced at $599, it offers a Snapdragon 7 Gen 1 processor and unique features like a 3D interactive pet and extensive customization options.

Another significant release was the Nubia Pad 3D II, a tablet capable of displaying 3D content without glasses through eye-tracking technology. This new version introduces 5G connectivity and incorporates “AI concepts” for enhanced functionality, including dual cameras for 3D content creation and an AI feature that converts 2D to 3D content.

Xiaomi 14 Smartphone

Xiaomi unveiled its flagship Xiaomi 14 Ultra at the Mobile World Congress. This model enhances its predecessor’s capabilities, offering an unparalleled display, a more durable construction. However, the high cost and specific target market of photography enthusiasts might limit its appeal. The device, starting at €1,499, with an optional Photography Kit for €199, introduces HyperOS, an interface designed to refine user experience.

Additionally, Xiaomi unveiled other devices including the Xiaomi Pad 6S Pro, Xiaomi Smart Band 8 Pro, and the Watch S3 with HyperOS, alongside the Xiaomi Watch 2 running on Google’s Wear OS. While these products won’t reach the US market, their launch in Europe demonstrates Xiaomi’s strategic global expansion.

Humane AI

The Humane Ai pin, introduced a few months ago, was on display at the MWC. The wearable device aims towards a future less dependent on smartphones. It is designed by former Apple employees, and aims for a screen-free existence, blending seamlessly into personal attire while offering sophisticated AI functionalities.

Priced at $699, with a $24 monthly subscription for connectivity and AI services, the Ai Pin operates through voice and gesture interactions, supporting up to 50 languages and adapting to local languages automatically. Humane emphasises privacy with features like an LED indicator for the camera and encrypted data management, marking the Ai Pin as an innovative step towards integrating AI into daily life without screens.

The post 10 Futuristic Gadgets Announced at MWC 2024 appeared first on Analytics India Magazine.

“I think the biggest transformation I’ve seen is not just OCI but Oracle itself becoming a lot more cloud- and operations-focused,” said Pradeep Vincent, chief technical architect, Oracle, in an exclusive interview with AIM, adding that operations are one of the biggest value propositions of the cloud.

Vincent joined the company a decade ago and is one of the founding members of Oracle Cloud Infrastructure (OCI). He has been involved in the design and implementation of the cloud infrastructure, responsible for the engineering architecture group, and plays a key role in driving the development of OCI.

Multi-Cloud Approach

Vincent believes that OCI is pretty different from the competitors out there. “Our goal is to make it easy for customers to use multiple clouds, period,” he said, explaining that a key part of this is their ‘distributed cloud strategy’, putting the cloud where customers want it.

“There are a few different ways in which we are going about it, one of them is multi-cloud, with Oracle Database@Azure being one. We’re super excited about that,” he said.

Oracle, last year, announced Oracle Database@Azure, which delivers Oracle database services running on OCI inside Azure datacenters and gives customers more flexibility in where they run their workloads.

“I worked on the engineering architecture behind the scenes for that. It’s truly impressive,” said Vincent, adding that they essentially took OCI itself, creating a small OCI site inside the Azure Data Centre and connecting it to OCI.

Simultaneously, it was wired directly to the Azure network. It is only a matter of time before OCI is integrated with GCP and AWS. “Cloud should be open,” said chief technology officer Larry Ellison.

OCI’s Architecture

Vincent is confident in the networking architecture of OCI. “I believe our super clusters are exceptionally powerful. Customers frequently tell us that when they try other cloud providers, including OCI, they observe a significant difference in networking technology,” he said.

He further explained that good networking facilitates proper utilisation of GPUs. “GPUs are very pricey. If you think about LLM training, many of them actually run like 40 to 60% utilisation of the GPUs,” he said, adding that although networking comes with a cost, it provides a 10x value in terms of GPU savings.

OCI Stands for Security

“OCI is secure by default, which means it embeds security features as a built-in aspect. It’s not like buying a product where you have to add an extra pack for security, here, it’s integrated from the beginning,” said Vincent.

OCI provides a variety of features to help you secure your network, such as subnet network filtering, firewalls, and security lists. Moreover, OCI offers a variety of data encryption options, both at rest and in transit. This helps to protect your data from unauthorized access, even if it is intercepted.

Vincent explained that Oracle doesn’t make a network public by default. Users have the option to make it public if they choose to. However, he cautioned that a common issue arises when customers unknowingly leak data into the public cloud by creating public buckets.

“Our security story is not just about infrastructure or apps but goes end-to-end. Advanced security functionalities like Identity & Access Management and Cloud Guard cut across most threats. And that’s a huge differentiator as far as this is concerned,” said Vincent.

Generative AI in OCI

Oracle recently embedded generative AI capabilities into the complete SaaS suite, which include applications like ERP, HCM, SCM, and CX. Additionally, OCI offers models from Cohere and Meta for various tasks such as writing, summarisation, analysis, and chat, without requiring extensive training from scratch.

“The way I look at generative AI is that it is a very novel and creative way to use data. But it’s not necessarily accurate on its own every single day,” said Vincent, adding that generative AI will enhance use cases but not necessarily replace all.

Speaking from Oracle’s perspective, he said the company is focusing on expanding data centres to meet the demand for generative AI services. “We offer many unique services, including cluster networks with support for remote direct memory access (RDMA),” said Vincent. He further mentioned that Oracle internally uses generative AI for customer support.

When asked about his motivation, he shared, “There’s a lot of customer problems to be solved, lots of innovation that’s happening. So I’m excited and privileged to be part of it.”

The post The Brain Behind Oracle Cloud appeared first on Analytics India Magazine.

Image by pch.vector on Freepik

If you want to become a skilled data scientist, you should know how to understand and analyze data. And for this statistics is important.

However, learning statistics can feel difficult, especially if you’re not from a math or computer science background. But don't worry. We’ve compiled a list of statistics courses—from introductory statistics to slightly more advanced concepts—which you can take for free.

You don't have to take all of these courses to become proficient in statistics for data science. So please feel free to check out the courses that you particularly find interesting. Let’s get started!

Note: You can audit all of the following courses for free on Coursera.

1. Introduction to Statistics

The Introduction to Statistics course from Stanford is a good first course in statistics. This course aims at teaching all the statistical thinking concepts that are necessary to understand and analyze data.

Here’s an overview of the course contents your is an overview of what the course covers:

• Introduction and descriptive statistics for exploring data
• Producing data and sampling
• Probability
• Normal approximation and binomial distribution
• Sampling distributions and the central limit theorem
• Regression
• Confidence intervals
• Tests of significance
• Resampling
• Analysis of categorical data
• One-Way Analysis of Variance (ANOVA)
• Multiple comparisons

Link: Introduction to Statistics

2. Basic Statistics

Basic Statistics from the University of Amsterdam is also another beginner-friendly statistics course. This course requires you to be familiar with R programming and covers the following topics:

• Exploring data
• Correlation and regression
• Probability and probability distribution
• Sampling distributions
• confidence intervals and significance tests

Link: Basic Statistics

3. Statistics for Data Science with Python

The Statistics for Data Science with Python is offered by IBM as part of the Data Science Fundamentals with Python and SQL specialization.

This course will teach you how to use Python to perform statistical tests and interpret the results of statistical analyses. The contents of this course are as follows:

• Basics of Python
• Introduction and descriptive statistics
• Data visualization
• Introduction to probability distributions
• Hypothesis testing
• Regression analysis

Link: Statistics for Data Science with Python

4. The Power of Statistics

The Power of Statistics is offered by Google as part of their Google Advanced Data Analytics Professional Certificate.

From summarizing datasets to conducting hypothesis tests and modeling data using probability distributions, this course also focuses on statistical analysis with Python. This course covers the following topics:

• Introduction to statistics
• Probability
• Sampling
• Confidence intervals
• Introduction to hypothesis testing

Link: The Power of Statistics

5. Statistics with Python

The Statistics with Python Specialization offered by the University of Michigan teaches you how to use Python for data visualization, statistical inference, and modeling. It also emphasizes the importance of connecting the business questions you need to answer to the relevant data analysis methods.

This is a three-course specialization that covers the required theory as well as Python programming assignments to help you apply all that you’ve learned. The courses in the specialization are as follows:

• Understanding and Visualizing Data with Python
• Inferential Statistical Analysis with Python
• Fitting Statistical Models to Data with Python

Link: Statistics with Python Specialization

Wrapping Up

And that's a wrap. We went over five courses that you can take for free to learn statistics and level up your data science skills.

Because most of these courses focus on programming and running statistical tests with Python as opposed to learning only theoretical concepts, I’m sure you’ll find plenty of opportunities to apply what you’ve learned. Happy learning, and keep coding!

Bala Priya C is a developer and technical writer from India. She likes working at the intersection of math, programming, data science, and content creation. Her areas of interest and expertise include DevOps, data science, and natural language processing. She enjoys reading, writing, coding, and coffee! Currently, she's working on learning and sharing her knowledge with the developer community by authoring tutorials, how-to guides, opinion pieces, and more.

More On This Topic

• 3 Free Statistics Courses for Data Science
• 25 Free Courses to Master Data Science, Data Engineering, Machine…
• 5 Free Courses to Master Data Science
• 5 Free Courses to Master Python for Data Science
• 5 Free Courses to Master Data Engineering
• 5 Free Courses to Master Machine Learning

Shorthills AI today announced that they have formed a strategic alliance with Databricks, to improve business operations through the integration of artificial intelligence (AI) and data analytics. With this partnership, the company looks to provide a powerful platform that simplifies working with AI and analytics. The Databricks platform helps businesses store, process, and analyse large volumes of data efficiently.

Shorthills AI was founded in 2018 by Pawan Prabhat and Paramdeep Singh, who come from strong technical backgrounds and a shared vision for leveraging AI to solve complex business challenges. The company provides a comprehensive solution to their clients from organising and cleaning up their messy data to machine learning and AI solutions.

The latest partnership benefits both companies and their clients. For Shorthills AI, collaborating with Databricks means accessing a robust platform that enhances their ability to handle big data and complex AI projects. This collaboration allows Shorthills AI to focus on developing tailored AI solutions for their clients, knowing they can rely on the strong data management and analytics capabilities of Databricks.

On the other hand, Databricks benefits from the partnership by leveraging Shorthills AI’s expertise in creating custom AI solutions. This expertise helps Databricks to further penetrate markets and industries where specialised knowledge is crucial for addressing specific business challenges. Together, Shorthills AI and Databricks aim to transform raw data into actionable insights and intelligent solutions.

For clients, this partnership means they can harness Shorthills AI’s deep understanding of AI and machine learning, combined with Databricks’ scalable and efficient data platform.

“We are thrilled to partner with Databricks, as this collaboration aligns with our mission to empower businesses through the strategic use of AI and data analytics,” said Paramdeep Singh, Co-founder, Shorthills AI.

Previously, Databricks has also partnered with major Indian companies such as Infosys, Wipro, Tata Consultancy Services (TCS), 3i Infotech, and Celebal Technologies. These partnerships aim to leverage Databricks’ data and AI platform across various industries. These collaborations highlight Databricks’ strategy to broaden its impact in India, a vital market for data and AI innovations. For businesses, this means access to advanced data analytics and AI expertise, supporting their operational and strategic goals.

The post Shorthills AI, Databricks Partners to Simplify AI & Analytics for Businesses appeared first on Analytics India Magazine.

In a historic move, the Union Cabinet has given its nod to three significant proposals for semiconductor plants, marking a monumental step towards bolstering India’s capabilities in the semiconductor industry. Among these proposals, the Tata Group and Taiwan’s Powerchip Semiconductor Manufacturing Corporation (PSMC) will spearhead the establishment of India’s first semiconductor fab in Dholera, Gujarat, at an estimated cost of Rs 91,000 crore.

Tata’s JV in Semiconductor Fabrication

Announcing this groundbreaking development, IT Minister Ashwini Vaishnaw revealed that the Tata joint venture (JV) is set to create India’s inaugural semiconductor fab with a noteworthy capacity of 50,000 wafers per month. The decision reflects a strategic move to strengthen India’s position in the semiconductor manufacturing landscape.

“Today the Prime Minister has taken an important decision to set up a semiconductor fab in the country. The first commercial semiconductor fab will be set up by Tata and Powerchip-Taiwan, whose plant will be in Dholera. After Micron semiconductor plant in Sanand, now a fab to come up in Dholera,” said Vaishnaw.

Assam and Gujarat to House Semiconductor Units

In addition to the Dholera project, the Cabinet has approved Tata Semiconductor Assembly Test’s semiconductor assembly and testing unit in Assam, representing a substantial investment of Rs 27,000 crore. This move underscores the government’s commitment to dispersing semiconductor capabilities across the country.

Furthermore, CG Power and Japan’s Renesas have received approval to establish a semiconductor plant in Gujarat’s Sanand, set to produce 15 million chips per day at an estimated cost of Rs 7,600 crore. This diversification aims to create regional hubs for semiconductor manufacturing, promoting decentralized growth in the sector.

India’s Semiconductor Ambitions

The recent greenlighting of these semiconductor projects aligns with the Ministry of IT’s earlier announcement that India will soon house two semiconductor fabrication plants, accompanied by several chip assembly and packaging units. These strategic initiatives are poised to impact India’s semiconductor landscape significantly, reducing dependence on imports and positioning the country as a formidable player in the global semiconductor industry.

These semiconductor projects come on the heels of a Rs 22,516-crore chip assembly plant being established by US-based memory chip maker Micron, signalling a comprehensive approach to developing a robust semiconductor ecosystem within the country.

The proposed investments, totalling an estimated Rs 1.26 lakh crore, reflect the government’s dedication to fostering innovation, technology, and self-reliance in critical sectors, ultimately positioning India as a global semiconductor manufacturing player.

The post Govt Approves India’s First Semiconductor Fab Led by Tata Group and PSMC appeared first on Analytics India Magazine.