Datatonic wins the Google Cloud Specialisation Partner of the Year Award

By

-

June 17, 2022

Datatonic wins Google Cloud Specialisation Partner

Datatonic, a Data + AI consulting firm on Google Cloud, has announced that it has won the Google Cloud Specialization Partner of the Year 2021 award for Machine Learning.

The data consultancy was awarded for its achievements in the Google Cloud ecosystem. It has been helping the joint customers enhance their Machine Learning (ML) functions with Machine Learning Operations (MLOps) and impact their business with transformational ML solutions.

Datatonic is a London-based data consultancy that assists companies in making better business-related decisions with the help of Modern Data Stack and MLOps. The services provided by the consultancy enable clients to deepen their understanding of consumers, unlock operational efficiencies, and increase competitive advantages. Datatonic builds cloud-native data foundations, employing high-impact analytics and machine learning use cases.

In the past years, Datatonic has built high-performing MLOps platforms for international clients across the media, telecommunications, and e-Commerce sectors by enabling them to leverage MLOps best practices throughout their teams.

The recently open-sourced MLOps Turbo Templates showcase Datatonic’s expertise in implementing MLOps solutions and Google Cloud’s technical excellence to help teams get started with MLOps. It was co-developed with Google Cloud’s Vertex AI Pipelines product team.

According to Nina Harding, Global Chief of Partner Programs and Strategy, Google Cloud, Google Cloud Specializations recognize partner excellence and proven customer satisfaction in a particular product industry or area. She added that based on their certified strong technical capabilities and customer success, Datatonic had been recognized as Specialization Partner of the Year for Machine Learning.

UK Govt launches Future of UK Defense Artificial Intelligence

By

Dipayan Mitra

-

June 17, 2022

The government of the United Kingdom announced a plan for the future of UK artificial intelligence defense technology. The new strategy was unveiled during the recently held London Tech Week AI Summit.

According to officials, a new Defense AI Center (DAIC) will be established to promote, empower, and innovate AI technologies throughout the UK Armed Forces with speed and ambition, based on the strategy and associated policy on ‘Ambitious, Safe, and Responsible’ AI usage.

The Defense AI Strategy outlines how the United Kingdom will prioritize research, development, and experimentation to transform the country’s Armed Forces capabilities through innovative concepts and cutting-edge technology. This new development will help the UK in delivering the most up-to-date equipment to the battlefield via effective, efficient, and dedicated routes.

Defense Procurement Minister, Jeremy Quin, said, “Future conflicts may be won or lost on the speed and efficacy of AI technology, and our approach to AI must be rapid, ambitious, and comprehensive.”

Quin further added that their new Defense AI Center (DAIC) and AI strategy would provide a dedicated center to advocate these technologies, working ethically with human judgments to keep the UK at the forefront of global security and responsible innovation. Soldiers on the front lines directed by smart systems relying on hours of detailed footage taken by a series of tiny drones are a few examples of the ideas that the government has.

Moreover, the government also released a policy on the “Ambitious, Safe, and Responsible” use of AI, which was developed in collaboration with the Center for Data Ethics and Innovation (CDEI) and included new efforts for AI in defense.

Head of Dstl’s Cyber and Information Systems Division, Dr. Paul Kealey, said, “AI has the potential to provide significant benefits across Defence from the back-office to the Front Line, and I’m delighted we are working with Kainos – a brand new supplier who will bring specialist expertise and experience as a leader in the civil world into defense.”

DALL-E Mini, an open alternative to DALL-E, goes viral

By

Disha Chopra

-

June 16, 2022

Boris Dayma designed an open alternative to DALL-E by the name of DALL-E Mini. As the name suggests, it is a “mini” version of the former. The DALL-E Mini is an image generator that employs artificial intelligence to generate a series of images depending on user-entered prompts. It can visualize most weird notions or events in a photo grid format.

Lmao I can’t believe how accurate this thing is#dallemini pic.twitter.com/YxQm92HYF9
— Bryan (@twerkforseitan) June 9, 2022

According to programmers and researchers, large language-based datasets can train a vast neural network to perform diverse text production tasks and generate high-resolution images. In January 2021, OpenAI developed DALL-E, a multimodal generative neural network to create images from text.

DALL-E uses the trained GPT-3 parameter version to create images using text-image pairs. It is the world’s first creative artificial intelligence model that can generate visuals from the text. In other words, it blends natural language understanding with the generation of realistic pictures.

I'm dying #dallemini #Godzilla pic.twitter.com/aggte4NH49
— T_Zanc (@tylerzanca) June 9, 2022

Following its success and mettle, OpenAI released DALL-E 2, an updated version in April 2022. This version incorporated higher resolution and lower latency. However, DALL-E is not open to the public. Thus, the DALL-E Mini is based on the same concept devised by OpenAI but is free and open to the public. Unfortunately, there is not much that DALL-E mini offers against DALL-E.

For everyone who asked for it, here's a Lisa Frank Xenomorph. #dallemini #dalle pic.twitter.com/ahBLsadcyI
— 💛🤍💜🖤 RED 💚🤍🖤 (@prettypinkpansy) June 8, 2022

Both DALL-E and DALL-E Mini share the same architecture. It is only during the training that the mini version enters a much larger dataset. DALL-E 2 works on a 3.5 billion parameter model and another 1.5 billion parameters to improve the resolution of produced images.

Boris Dayma, the creator of DALL-E Mini, said, “The model is trained by looking at millions of images from the internet with their associated captions. Over time, it learns how to draw an image from a text prompt. Some of the concepts are learned from memory as it may have seen similar images. However, it can also learn how to create unique images that don’t exist such as “the Eiffel tower is landing on the moon” by combining multiple concepts together.”

NXP announces MCX general-purpose Arm Cortex-M MCU family with 30x faster ML performance

By

Sahil Pawar

-

June 16, 2022

NXP announces MCX general-purpose Arm Cortex-M MCU

NXP has announced the launch of its new MCX general-purpose Arm Cortex-M MCU family, designed specifically for advanced industrial and IoT edge computing. The NXP neural processing unit (NPU) can deliver over 30 times higher machine learning (ML) performance than running the AI inference tasks on an Arm Cortex-M33 core.

The new MCX portfolio is built upon the previous NXP LPC and Kinetis microcontroller families and thus does not replace them. The portfolio aims to enhance performance and security for numerous applications, including machine learning, voice, wireless, analog, motor control, and more.

The new MCX general-purpose Arm Cortex-M MCU family will be available in four series:

MCX N Advanced series is designed for secure and ingenious applications. With a frequency of 150 MHz to 250 MHz, it has DSP, a Neural processing unit (NPU) for real-time inference, and an EdgeLock Secure Subsystem.

MCX A Essential series is optimized to deliver critical functionality for applications like motor control with a frequency of 48 MHz to 96 MHz. Besides being optimized for cost-constrained applications, it has built-in timers, a single-pin power supply, and a low pin count.

MCX W Wireless series has an MCU with low-power narrow band connectivity, Bluetooth Low Energy, and a frequency of 32 to 150 MHz. It possesses on-chip integration to reduce BOM costs.

MCX L Ultra-Low Power series is designed for power-critical applications. It has a frequency of 50 to 100 MHz with an optional 50% boost. It possesses the lowest active power and leakage of the MCX family.

The MCX MCU will be supported by the MCUXpresso software and tools similar to the Kinetis and LPC MCUs. The eIQ ML SW development environment will also help it for AI inference on the MCX N Advanced series with an NPU and other MCX microcontrollers using the Cortex-M core.

Detailed specifications and documentation are not provided by NXP yet. NXP will showcase the MCX W series next week at the Embedded World 2022.

The company has emphasized that the MCX portfolio is not replacing Kinetis and LPC portfolios. MCX is a new NXP portfolio that was purpose-built based on NXP’s expertise in microcontrollers. The company added that Kinetis and LPC portfolios have been experiencing high demand and are still industry-relevant devices.

H2O.ai Expands Snowflake Partnership enabling AI Transformations for Customers

By

Dipayan Mitra

-

June 16, 2022

Open-source machine learning platform H2O.ai expands its partnership with Snowflake to enable artificial intelligence (AI) transformation for its customers.

By seamlessly connecting data and machine learning, the company demonstrated a unique set of capabilities and use cases that enable rich insights.

According to the company, Snowflake and H2O.ai have developed a native integration that allows users to access all of H2O.ai’s advanced machine learning capabilities directly from their Snowflake environment.

This integration helps innumerable customers across the globe to innovate with AI. Users can employ H2O.ai machine learning capabilities for near real-time analysis from within the Snowflake Data Cloud, thanks to several pre-built interfaces.

Head of Technology Alliances at Snowflake, Tarik Dwiek, said, “Our partnership with H2O.ai can help optimize the supply chain across multiple industries including manufacturing, telecom, banking, and retail.”

He further added that users could decrease risks, enhance customer experiences, drive growth, and increase efficiency by using H2O AutoML to produce, operate, and innovate for their customers and partners.

The United States-based artificial intelligence company H2O.ai was founded by Cliff Click and Sri Satish Ambati in 2012. The company specializes in providing solutions to solve complex business problems and accelerate the discovery of new ideas. To date, H2O.ai has raised over $251 million over eight funding rounds from investors like Commonwealth Bank of Australia, Goldman Sachs, Celesta Capital, Crane Ventures Partner, and others.

Sri Ambati said, “H2O AI Cloud is democratizing AI on Snowflake’s Data Cloud, helping our customers personalize their offerings and bring efficient flows into their business.” Ambati also mentioned that the strong multi-cloud integration of H2O AI engines and AI App Stores promotes consumption on Snowflake Data Cloud, lowering costs for customers.

A few months earlier, H2O.ai also announced the launch of its new deep learning training engine named H2O Hydrogen Torch. The unique training engine allows companies of all sizes in any industry to make impeccable images, videos, and natural language processing (NLP) models.

Iterative announces machine learning-based extension for Microsoft Visual Studio Code

By

Sahil Pawar

-

June 16, 2022

ML-based extension for Visual Studio Code

Iterative, the MLOps company working toward strategizing the workflow of data scientists and machine learning (ML) engineers, has announced a free extension for Visual Studio Code (VS Code). The extension is a source-code editor developed by Microsoft for machine learning model development and experiment tracking.

The extension will simplify machine-learning model development workflows for data scientists and meet the ML modelers working there. The extension eliminates the need for expensive SaaS solutions for experiment tracking by turning VS Code into a native machine learning (ML) experimentation tool built for developers.

VS Code is a coding editor that allows users to initiate coding in any programming language quickly. The DVC extension for Visual Studio Code will allow users from all technical backgrounds to create, visualize, compare, and reproduce machine learning experiments. The extension makes experiments easily reproducible through Git and Iterative’s DVC. This is unlike traditional experiment tracking tools that only stream metrics.

The extension enhances the existing VS Code UX with features using Source Control view, Command Palette, File Tree explorer, and custom in-editor web views. These features aid data scientists in model development and experimentation workflows. Through this extension, users can run and reproduce experiments, pull and push versioned data, and view metrics and tables.

Beyond the tracking of ML models, hyperparameters, and metrics, this extension makes ML experiments reproducible by tracking data changes and source code. According to Dmitry Petrov, CEO of Iterative, the company’s experiment versioning technology implemented in DVC last year makes this reproducibility possible.

Additionally, the extension provides resource tracking for data scientists to see which data sets and models have changed. It also allows the exploration of all project or model files. Other features of the extension include live tracking of metrics, native plot visualization, and

cloud-agnostic data versioning and management.

Meta Pharmaceuticals Raises $15M to Make Autoimmune Drugs with AI, New Immuno-Metabolism Tech

By

Disha Chopra

-

June 16, 2022

Shenzhen-based biotech startup Meta Pharmaceuticals lands its initial investment of $15M. The company is about ten months old and is an XtalPi AI-assisted project. XtalPi is an industry incumbent in pharmaceutical technology.

With the investment, Meta aims to make drugs for autoimmune diseases with the assistance of AI. The acquisition is a total of seed and pre-A funding rounds that included Forcefield Ventures, IMO Venture, Tiantu Capital, and XtalPi itself. Both XtalPi and Meta have a complementary role in translating new drug targets into patents. Both companies are ambitious to capture the market with AI-based marketable drugs.

Meta’s technology and work fall under immuno-metabolism to study the relationship between immunology and metabolism. Drugs belonging to the category are purposed to regulate the immune system. As of now, XtalPi will cater to the Initial stages of the discovery of therapeutic targets and their molecular design. Once the drug reaches the pre-clinical phase, Meta shall take over to facilitate further development, new drug filing, and trials.

It will use public data for initial training and plans to collect samples and patient data in collaboration with hospitals in China further down the process. An XtalPi spokesperson said, “Our pipeline has the potential to be used for treating a wide range of autoimmune problems, cancer and age-associated diseases. It’s still too early to share the specifics just yet or limit ourselves to one or two existing drugs and indications.”

Indeed, Meta plans to take its drug to the global market. Currently, the startup is receiving support from Shenzen and Hong Kong in its infancy. Many government-led initiatives like tax exemptions are being provided to Meta and other similar startups.

Tabnine, an AI Code Completion Startup, raises $15.5M to expand its AI

By

Disha Chopra

-

June 16, 2022

Tabnine will use its $15.5M funding to create an AI-powered assistant that autocompleted code for software developers. Qualcomm Ventures, OurCrowd, and Samsung NEXT Ventures co-led the funding round with Kholsa Ventures and Headline Ventures. The proceeds of this investment will improve developer experience, add new capabilities, and strengthen Tabnine’s offerings.

Tabnine was founded by Dror Weiss and Eran Yahav in 2012 as Codota, named later Tabnine around 2017. It employs AI to complete chunks of code after knowing the purpose. It uses trained tapping algorithms to understand the code models and then attempts to learn the developer’s practices.

“Based on our previous work on code analysis and simulation, we realized that with the vast amount of commonality and standard patterns in code, it was inevitable that AI will be a critical part in the dev process. We set out and pioneered the AI code assistant category”, said Dror Weiss, CEO, and Co-founder of Tabnine.

Tabnine provides improvements on every keystroke. Developers can also have full-line recommendations within the development environments like VSCode, IntelliJ, Eclipse, Android Studio, and Webstorm. The platform provides ‘code-native’ AI models trained in specific programming languages. It supports Java, C++, PHP, Go, C#, Python, Ruby, JavaScript, Rust, Swift, TypeScript, Haskell, OCaml, Scala, Kotlin, Perl, SQL, CSS, HTML, and Bash.

Weiss also claims that Tabnine’s approach enables it to learn the patterns in code efficiently. He said, “Our models give customers the flexibility to run Tabnine either on our cloud or on their network, and the ability to train custom AI models that capture the specific patterns in their repositories.”

Metropolis Technologies raises $167 million in Series B Funding Round

By

Dipayan Mitra

-

June 16, 2022

Mobility commerce platform Metropolis Technologies raises $167 million in its recently held series B funding round led by 3L Capital and Assembly Ventures.

Several other investors, including Dragoneer Investment Group, Eldridge, Silver Lake Waterman, and UP Partners, also participated in the company’s latest funding round.

Metropolis has witnessed 28x user growth and is currently active in over sixty cities even after facing the difficulties that the COVID-19 pandemic brought.

The company enables drive-in and drive-out payment for millions, runs over 600 parking lots and garages, and collaborates with thousands of grocery stores, coffee shops, and other neighborhood retailers.

Co-founder and CEO of Metropolis Technologies, Alex Israel, said, “By bringing urban mobility infrastructure online, we are knitting disconnected parts of the everyday journey into a remarkable experience.”

He further added that the seamless transactions that they take for granted in eCommerce are mostly absent in real life, but Metropolis’ computer vision technology is changing that one commute, one-morning coffee, and one supermarket run at a time.

United States-based artificial intelligence and computer vision startup Metropolis was founded by Alexander Israel, Courtney Fukuda, Peter Fisher, and Travis Kell in 2017. The company specializes in developing technologies to modernize parking and empower the future of mobility. Metropolis’s platform also connects users to local business specials and discounts, generating new economic prospects for thousands of local companies.

The parking activity data collected by Metropolis can be used to assist users in finding parking spots faster—a potentially important service in a car-centric metropolis like Los Angeles. To date, Metropolis has raised $228 million from multiple investors over three funding rounds.

Activision’s subsidiary King acquires Peltarion

By

Dipayan Mitra

-

June 16, 2022

Interactive entertainment company and Activision’s subsidiary King acquires real-world artificial intelligence applications developing firm Peltarion. However, neither company provided any information regarding the valuation of this acquisition deal.

This new acquisition of Peltarion will help King boost up the contemporary use of AI and systems, getting to know the era in King’s recreation platform. According to the company, this acquisition is a strategic step toward the company’s direction.

King intends to continue building top-tier AI and machine learning skills and teams, allowing a new generation of creative game design, development, and live operations capabilities, along with becoming a center for the world’s best talent in-game AI.

“We are proud to announce the acquisition of Peltarion. Machine learning technology is moving quickly, and by increasing our investments in this field, we expect to deliver even more creative content to our 250 million monthly players* (Q1 2022) across the globe,” said Tjodolf Sommestad, President of King.

Sommestad further added that they believe that the talented Peltarion team, combined with the powerful technology they have developed, will enable them to serve their players even better with more engaging games and content.

Stockholm-based artificial intelligence company Pelrarion was founded by Luka Crnkovic-Friis and Måns Erlandsson in 2004. The company specializes in offering a graphical cloud platform for collaboratively designing, managing, and deploying deep learning systems at scale. NASA, Tesla, iZettle, General Electric, Dell, BMW, Deutsche Bank, Lloyds Banking Group, and the universities of Harvard, MIT, and Oxford have all leveraged Peltarion’s AI technology.

“The scale and reach of King, with iconic franchises like Candy Crush Saga, is a great match for our technology. The opportunities with AI seem to be endless, and we cannot wait to unlock the potential of working together,” said Luka Crnkovic-Friis, Peltarion CEO.