Vetiver, an MLOps framework for R and Python, was released by RStudio. It provides fluent tooling to share, version, deploy, and monitor trained models. There are many open-source tools that data scientists love to use to prepare data for training. However, they are not fluent. Data scientists and developers acknowledge the lack of fluent open-source tooling for MLOps processes like fitting a model in production, monitoring performance, etc.
Vetiver for MLOps enables developers to use open-source tools for exploratory data analysis and model tuning. Currently, it supports all models trained via PyTorch, tiny models, caret, scikit-learn, mlr3, XGBoost, ranger, lm(), and glm(). The framework is open to incorporating several other modeling frameworks in further developments.
Vetiver is specially designed for data science teams that use Python and/or R. Vetiver’s approach considers “bilingual” data science projects so that the developers may use either (or use both together).
The framework provides a flexible and extensible first deployment experience. RStudio believes that the tools that beginners use are also the ones that help data practitioners as they age advance. This is why, to avoid typical failure modes, functions in the vetiver framework accommodate both recording and validating the model’s input prototype. Vetiver functions tackle both simultaneously. To deter failures in deployment, Vetiver also enables the user to predict from a remote API endpoint so that a deployed model is treated similarly to a local R/Python one.
Vetiver is open for up-gradation. If you have any concerns or ideas, you can join the discussion on RStudio Community.
China launches the world’s first artificial intelligence (AI)-controlled unmanned carrier capable of launching multiple drones, reports Business Insider.
Although the carrier is designated as a marine research vessel, some experts believe it has the potential to be deployed as a military vessel.
Developed by CSSC Huangpu Wenchong Shipping, the unmanned carrier named Zhu Hai Yun is an autonomous ship that can carry dozens of air, sea, and underwater drones outfitted with various observation devices.
The ship was initially unveiled in May, according to the South China Morning Post, and is expected to be delivered by the end of 2022 after completing sea testing.
“From a purely science standpoint, which is the angle China is promoting, we could see Chinese drones (both surface and subsurface, and launched from the Zhu Hai Yun) contributing to disaster mitigation, environmental monitoring, etc.,” said Matthew Funaiole, senior fellow of China Power Project at the Center for Strategic and International Studies. He further added that the most immediate benefit to China is likely data collection.
Zhu Hai Yun uses the Intelligent Mobile Ocean Stereo Observing System, created by the Southern Marine Science and Engineering Guangdong Laboratory. It can be controlled remotely and has a top speed of 18 knots or approximately 20 miles per hour.
In 2021, Chen Dake, the laboratory’s director, told Science and Technology Daily that the ship is a new “marine species” that will revolutionize ocean surveillance.
China has recently been making several efforts toward the advancement of artificial intelligence. China is now also working on a new project to build an AI-powered 3D printed hydroelectric dam in Tibet. The dam will be constructed in just a period of two years on the Tibetan plateau.
International Business Machines (IBM) has announced the launch of a new Automation Innovation Centre at the IBM India Software Lab in Kochi. The center is expected to be fully operational by quarter 3 of 2022.
The Automation Innovation Centre will allow IBM and its ecosystem partners to build automation solutions through their product design, support, and engineering lifecycle. This function will help the clients in the areas of business automation, integration, and AIOps.
The IBM center will collaborate with partners and clients to co-create and foster innovation. It will also work on promoting an outside-in view while designing new products at speed and enhancing customer experience simultaneously. IBM is focused on acquiring local talent to bolster its growth in the region of the center.
With an intent to commercialize the increasing demand for automation softwares in the IT sector, IBM is all set to collaborate with Wipro and Tata Consultancy Services (TCS). The engineering teams of TCS and Wipro will co-locate to Automation Innovation Centre to co-innovate and co-create solutions in AI-powered automation.
Similarly, to help address the deficiency of AI-related skills and expertise in the IT sector, IBM has collaborated with the Indian Institute of Information Technology (IIIT) Kottayam. The initiative will encourage curriculum enhancement and equip students with the latest in-demand skills necessary to thrive in the technology sector.
To curate a two-week online course designated to assist business professionals with AI, Deloitte collaborates with the Business School of the University of Sydney. The new course demystifies AI, a rapidly-growing tech, for industry experts. There is a lack of proper AI-integrated business practices, and it often becomes a barrier to adopting new technologies. The course aims to counter the absence of AI-fluency among decision-makers.
As per Dr. Sandra Peter, Director of Sydney Business Insights, “AI fluency is not about being able to code or program an AI. It’s about knowing what AI can achieve and how it works, how it can fit into your organization, and how you can develop an AI strategy that creates value in a workplace. We want AI fluency to be rewarding for its people.”
The partnership is not only limited to the course but also aims to grow Australia’s investment in AI and its potential. Dr. Kellie Nutall, Deloitte Australia’s Artificial Intelligence Lead, said, “AI technologies are still relatively new but highly disruptive for business leaders. Our clients have told us they’re looking for these kinds of education solutions, and we know that one key blocker to successfully scaling AI in Australia is fluency.”
The first AI fluency sprint commences on August 12, 2022. It will focus on real-world case studies related to AI. Dr. Peter will accompany Professor Kai Riemer, Information Systems and Organisations, to deliver this lecture.
The strategic collaboration was also welcomed by the Dean of the University of Sydney Business School, Prof Greg Whitwell. He expressed, “The Business School is proud of our strong relationships with industry. Our partnership with Deloitte is another fine example of how we re-imagine the future of business education to deliver real benefit and enhance the skills of ambitious leaders through continuous learning.”
Intel‘s Sapphire Rapids, a next-generation Xeon processor, is expected to make the chipmaker giant more competitive after falling behind AMD in terms of technology over the years.
With Sapphire Rapids on its way, Intel hopes it will beat Genoa, AMD’s next-generation Epyc chip, by incorporating industry-first technologies such as Compute Express Link, DDR5, and PCIe Gen 5.
Sandra Rivera, executive vice president and general manager of Intel’s data center and AI group, has confirmed a further delay in the rollout of the Xeon processor.
Despite the delay in launch, the x86 giant Intel has already bagged Nvidia as a prospective customer for its Xeon Scalable next-generation CPU. At BofA Securities 2022 Global Technology Conference, Nvidia’s CEO announced that the GPU giant would use Sapphire Rapids for its flagship DGX H100 system instead of AMD’s upcoming Genoa chip.
While explaining the delay, Rivera said that the delay is not the result of any issues with the advanced 10nm node of Sapphire Rapids’. Last year, Intel had renamed the 10nm node of its Xeno CPU to ‘Intel 7’. This was to convey the message that 10nm nodes’ manufacturing process is the same as the 7nm nodes by rival companies TSMC and Samsung. Rivera pointed to Intel’s Alder Lake chips developed for PCs and laptops as proof of Intel 7’s well-being.
Rivera emphasized that Intel’s Sapphire Rapids will be a leading product once it is launched. She also explained that the popularity of the processor might be more short-lived than expected as AMD’s Genoa chip will launch later this year, right after Sapphire Rapids.
In April, the Department for Promotion of Industry and Internal Trade introduced Open Network For Digital Commerce (ONDC) as an alternative to dominant eCommerce global giants like Amazon and Walmart. It’s a non-profit company that will display products and services from all eCommerce platform participants across the network. For example, if platforms like Flipkart and Amazon sign up for the ONDC platform, a user searching for a Smart TV will be able to see products available on both platforms.
ONDC has received about ₹150 crores in funding. It aims to utilize open specifications and open network protocols for promoting open networks. The beneficiaries will be small, micro, and medium enterprises, hotels, retail stores, restaurants, and delivery partners. It’s a first-of-its-kind initiative that aspires to democratize eCommerce. The platform aims to raise eCommerce penetration to 25% of India’s consumer purchases in the next two years, from the current 8% in a country of 1.35 billion people.
Unlike the current platform-centric digital commerce platforms, ONDC will work on a network-centric model. Buyer and seller-side applications that will be connected to ONDC’s open network will allow transactions between the buyer and seller. All three platforms in the open network will be interconnected: mobility, eCommerce, and delivery. There will be several buyers and provider-side applications at the customer and seller’s end.
The seller-side applications will receive buyers’ requests, used to publish their catalogs, and fulfill buyers’ orders. The buyer site application will help customers to search for products or services from multiple participants. Initially, there will be a single gateway to start the application, but it will be later expanded to more gateways. Through this common and open network, buyers will be able to access all the goods and services provided by multiple sellers on all connected seller-side applications. ONDC is not a platform or application, instead, it’s an open network that eliminates the need for a central intermediary.
Data Security and open-network
For data privacy concerns, ONDC will not store or view transaction data. All the policies regarding the exchange of data will comply with the Information Technology Act, 2000 and efforts will be made to comply with the emerging Personal Data Protection Bill. There will be provisions to protect the User’s Personally Identifiable Information (PII) and seller data critical to trade (i.e., competitive data) from third-party access.
Challenges faced by ONDC
ONDC aims to tap millions of small businesses that lack technical expertise. The government will need to run an awareness campaign to bring small businesses to the eCommerce channels. The Confederation of All India Traders (CAIT), a group representing 80 million small businesses, said that such businesses lack resources and order volumes to match the discounts offered by big eCommerce marketplaces such as Amazon and Flipkart. Another challenge would be establishing technology and easy-to-navigate platforms that lure merchants and customers.
UPI
India has led various population-scale initiatives to democratize markets — Unified Payment Interface (UPI) or Goods and Services Tax Network (GSTN) or the Unique Identification Authority of India (UIDAI). The National Payments Corporation of India developed UPI to simplify mobile banking and promote digital payments. NCPI offers a backend solution on which other platforms could offer UPI and a front-end solution as well. Various Platforms such as Phonepe, GooglePay, and Paytm offer transactions via UPI where every transaction is clocked twice – by remitting and receiving bank. ONDC is being called the ‘UPI of eCommerce.’
Dr. Hitesh Bhatt, director at Retailers Association of India, said, “The open network for digital commerce is a market-led initiative aimed at enabling interoperability in digital commerce based on the principles of openness. It aims to enable unbundling, democratizing, and unlocking value for all parts of digital commerce” at the ONDC Masterclass held in December 2021.
Small businesses and traders have always found it difficult to match eCommerce platforms’ deals and discounts. This platform aims to eliminate digital monopolies and help SMBs, local sellers, and startups sell online. The Competition Commission of India (CCI) has launched multiple investigations against Amazon and Walmart-backed Flipkart regarding complaints that these retailers promote alpha sellers. The Indian government is creating a level playing field with ONDC that will help businesses of all sizes to display and sell their products and services.
Will ONDC be successful?
ONDC will follow a network-centric model where sellers and customers will be transacting through an open network regardless of the application or platform they use. What’s unclear in the ONDC strategy paper is how businesses or Kirana stores that haven’t started selling online will be able to directly benefit from the open network. Since the seller-side applications will be registering to participate in the open network, SMBs will still need to register as sellers with eCommerce platforms to sell. ONDC will bring a power shift in the eCommerce industry, but its successful execution depends on how efficiently government can make digital interactions easier. Unlike UPI, which requires little digital interaction, Kirana stores and local supermarkets will have to familiarize themselves with digital platforms, which will take time.
Artificial intelligence-powered video dubbing startup Papercup raises $20 million in its recently held series A funding round from Comcast-owned Sky.
Other participants include Octopus Ventures, Local Globe, Sands Capital, Sky and Guardian Media Ventures, Entrepreneur First, and BDMI.
Papercup’s current investors include William Tunstall-Pedoe, whose team developed Amazon’s Alexa, Zoubin Ghahramani, senior research director at Google Brain, and former Uber top scientist.
According to the company, the freshly raised funds will assist Papercup in its quest to make the world’s video content available in any language.
Fund manager at Octopus Ventures, Zoe Reich, said, “Papercup’s use of AI to provide affordable, high-quality dubbing can unlock that content for audiences around the world and, in doing so, drive a 100-fold expansion in the video and audio translation market with Papercup at the forefront.”
Papercup is also planning to use the funds to expand its research around expressive voices, expand into new languages, and scale our offering in markets. This capital will enable the firm to expand its promising research and enter new content categories.
United Kingdom-based artificial intelligence company Papercup was founded by Jesse Shemen and Jiameng Gao in 2017. The startup specializes in offering a platform that employs machine learning to translate the voice track on videos, allowing any creator to access a global audience of seven billion people. Papercup’s platform is very effective as it allows media organizations, content creators, education providers, and multinational businesses to increase the value of their work by making it available to a worldwide audience.
Papercup CEO Jesse Sheme said, “People retain up to 70 percent more information when watching videos dubbed in their native language.” Jesse further added that they could tackle all types of material with truly expressive cross-lingual AI dubbing, making video and audio more accessible and entertaining for everyone.
In its latest announcement, Renesas Electronics Corporation, a premier manufacturer of advanced semiconductors, announced its collaboration with Reality AI, a company that offers embedded AI solutions. The partnership will entail Renesas’s acquisition of Reality AI in all-cash transactions. It will enhance the former’s endpoint AI services by providing more efficiency to make their products AIoT ready.
With the evolution of endpoint-protected workload environments, the world cannot be more connected than it is today. This makes embedding AI into products inevitable. Renesas has been offering such workload environments that are equipped to embed AI in highly secured Microprocessors (MPUs) and Microcontrollers (MCUs).
The acquisition of Reality Analytics’ AI tools will allow Renesas to improve its in-house capabilities. Renesas will be able to provide more comprehensive and optimized endpoint solutions for both software and hardware.
Besides Reality AI’s range of embedded tools, it also provides TinyML solutions catering to advanced non-visual sensing. Their fast-delivering ML algorithms are capable of fitting on the smallest MCUs. Renesas’s wide range of MCU portfolios can combine these technologies to deliver top-of-the-line AI.
Reality AI Tools, a Reality AI flagship specifically designed to provide analytics from sensory data, will assist in enhancing Renesas’s signal processing capabilities.
The acquisition will also bring a brand-new AIoT center-of-excellence to existence. Experts from Reality AI will come together at the facility in Maryland and form a software development base to address the requirements of customers eager to work with AI.
Hidetoshi Shibata, CEO and President of Renesas, remarked, “The addition of Reality AI’s AI solutions to our existing embedded AI portfolios will further solidify our position as a leading AIoT solution provider.”
Nala Robotics, an AI technology company, has announced the launch of Pizzaiola. It is a fully autonomous pizzeria robot that uses artificial intelligence (AI) and machine learning (ML) to replicate pizzeria-style foods precisely.
Equipped with natural language processing features, the robotic system can respond to voice menu orders and operational commands, thus saving customers’ time and drastically improving productivity. Additionally, the system checks more than 1,200 parameters every microsecond, from robot field of vision to food quality and point-of-sales. This thorough check ensures safety and enhanced productivity.
Pizzaiola can cook up to 50 pizzas an hour. It also offers 35 choices of toppings and cheeses, along with five doughs and four types of sauces. The robot can also prepare various kinds of salads, burgers, wings, and pasta.
Additionally, the modular kitchen setup in Pizzaiola allows customers to customize their orders by selecting from various ovens, fryers, and grills. The machine consists of traditional conveyor-style and brick-style pizza ovens.
After taking orders, Pizzaiola’s robotic arms knead the dough and add the toppings before sending it to the oven. After the cooking, the robot slices the pizza and packs it into boxes ranging from 8-18-inch in size.
Ajay Sunkara, co-founder and CEO of Nala Robotics, said that Pizzaiola could provide the same output as two full-time workers. He added that it is also capable of working 24-7-365 and can yield an excellent return on investment in a couple of years. Pizzaiola is now available as a restaurant-as-a-service with a monthly leasing option.
Jerome Presenti, Meta‘s vice-president of AI, is leaving after four years at the company amid a reorganization of its artificial intelligence group. The announcement comes a day after Sheryl Sandberg announced that she is leaving the position of Meta’s COO later this year.
The company said that Presenti would depart in mid-June after helping Meta through the early stages of the crucial AI transition. Meta is integrating its AI teams across various product groups instead of having the AI function as a centralized organization.
To continue advancing AI technology and community, Meta will convene a cross-functional AI leadership team led by Joelle Pineau, director of Meta AI research labs. With the new team structure, Meta said that it is all set to push the boundaries of what AI can achieve and use it to create new products for people worldwide.
The AI research team led by Yann LeCun, Meta’s chief AI scientist, will shift to Bosworth’s Reality Labs. This move will further consolidate resources in the AR/VR division, which has more than 17,000 employees.
According to Meta, AI is a critical component necessary to build the future hardware to define its vision of the metaverse. The company aims to take one step closer to the idea by reorganizing the AI group.
