Qualcomm and Google announce Partnership on Neural Architecture Search (NAS)

By

-

December 3, 2021

Qualcomm Google Partnership, Google Vertex AI NAS, Qualcomm Snapdragon — Image Credit: Analytics Drift Design Team

Chipmaker Qualcomm has teamed with Google to help expand its neural network technology across platforms. The firm will now use Google Cloud’s Vertex AI Neural Architecture Search (NAS) Services, which will first be available on the Snapdragon 8 Gen 1 platform — recently launched at the annual Qualcomm Snapdragon Tech Summit.

The Snapdragon 8 Gen 1 Mobile Platform is Qualcomm’s most powerful mobile platform to date, and it is the first in the company’s line-up to use a new branding approach that avoids the triple-digit naming tradition of its predecessors. From the cutting-edge process used to build the chips to its updated CPU, GPU, and AI processing engines, to its extensive camera and imaging technologies, and its comprehensive array of wireless connectivity options, Snapdragon 8 Gen 1 boasts significant advancements in virtually every aspect of the platform.

Qualcomm's new Snapdragon 8 Gen 1 chip is here to power the Android flagships of 2022 - The Verge — Image Source: Qualcomm

The Snapdragon 8 Gen 1 will be built on cutting-edge 4nm technology and include a combination of Arm CPU cores — a total of eight. A single high-performance Prime Cortex-X2 core (up to 3GHz), three Cortex A71 Performance cores (up to 2.5GHz), and four Cortex A51 Efficiency cores make up the upgraded Kyro CPU complex (up to 1.8GHz). The Prime core is utilized for threads that require the most priority (and performance), while the Performance cores do the remainder of the heavy lifting, with the Efficiency cores providing support for less-demanding background operations.

Snapdragon 8 Gen 1 unveiled with new ARMv9 CPU cores, new Adreno GPU architecture - GSMArena.com news — Image Source: Qualcomm

Through Snapdragon 8 Gen 1, Qualcomm Technologies will offer high precision AI with low latency to low-power devices like IoT, medical images, automobiles, and mobile devices using Google Cloud’s Vertex AI NAS, while maintaining memory and energy efficiency.

While the NAS will be accessible on Qualcomm’s new flagship Gen 1 mobile platform, it will gradually be rolled out throughout the Qualcomm range.

According to Qualcomm, NAS will be used to “accelerate neural network development and differentiation” for Snapdragon mobile, ACPC, XR, the Snapdragon Ride automotive platform, and IoT activities once it is integrated with 7th Gen Qualcomm’s Artificial Intelligence (AI) Engine.

Thanks to an improved Qualcomm Hexagon processor with double the shared memory and a tensor accelerator that’s twice as fast, the 7th Gen Qualcomm Artificial Intelligence (AI) Engine on-board is supposedly 4X quicker than its predecessor Snapdragon 888+ 5G. A Qualcomm Sensing Hub 3rd Generation is also included in the design, which handles the always-on sensors and consumes less power than its predecessors. Qualcomm claims a 1.7X gain in battery efficiency in addition to its massive AI performance boost.

Qualcomm claims that the Snapdragon 8 Gen 1’s Adreno GPU renders graphics 30% quicker than the Snapdragon 888 in this generation. This generation also outperforms the Snapdragon 888 in terms of power efficiency by a stunning 25%.

For developers, Google Cloud Vertex AI NAS will be included in the chipmaker’s Neural Processing SDK, operating on the Qualcomm AI Engine. Platforms that use the AI Engine will be able to benefit from “optimizations and performance enhancements,” adds Qualcomm. This will also allow the company to build and optimize new AI models in weeks rather than months.

Vertex AI Neural Architecture Search was launched by Google Cloud in May as a single platform for designing, deploying, and maintaining AI models. Vertex AI requires over 80% fewer lines of code to train models than existing platforms. Google claims it is the same framework that is used internally to power Google, with capabilities ranging from computer vision to language and structured data.

While Vertex AI is an assortment of various tools, Qualcomm emphasized on the Neural Architecture Search. Its goal, as the name suggests, is to improve AI models. NAS allows data scientists to tune a model for specific hardware without having to train it manually. Based on the use case, they can also impose limitations on the model’s size or other characteristics.

“The ability to utilize Google’s NAS technology to create and optimize new AI models in a condensed time frame is a game changer for our business,” said Ziad Asghar, Vice President of product management at Qualcomm. Ziad added, “We are happy to be the first chipset company to work with Google Cloud on NAS and eager to roll out this technology to further our momentum in connecting the intelligent edge.”

BigBear.ai secures First Position in NAVWAR’s AI Prize Challenge

By

Dipayan Mitra

-

December 3, 2021

Artificial intelligence and machine learning firm BigBear.ai secures the first position in the Naval Information Warfare Systems Command’s (NAVWAR) artificial intelligence and Networks Advanced Naval Technology Exercise challenge.

L3 Harris, one of the global leaders in developing avionics, tactical communications, and geospatial systems, secures the second position. The contest was arranged to address the gap between current and future warfare techniques and technologies.

The prize challenge was organized to support the US Navy’s Overmatch initiative, which focuses on modernizing naval warfare using artificial intelligence-powered weapons and sensors in a Naval Operational Architecture.

Chief Technology Officer of BigBear.ai, Brian Frutchey, said, “This challenge allowed us to demonstrate how our automated course of action assessment AI can assist the Navy in enabling warfighters to make critical decisions quickly in operationally relevant maritime environments.”

He further added that the company feels proud to participate in the competition to support Naval initiatives. The competition had numerous participants across various fields, including commercial, government, and academics.

Science and Technology director at NAVWAR, Carly Jackson, said, “The participants had less than three months but the results we are seeing are quite compelling. By quickly leveraging the lab infrastructure and expertise resident across the Naval Research and Development Establishment, this new type of digital platform-powered ANTX enables us to identify and field technologies, components, or algorithms at the speed of the threat.”

She also mentioned that the participating teams sought to motivate industry-leading companies to bring new innovation in their platforms and architectures.

Earlier this month, BigBear.ai also partnered with Palantir, a leading data analysis and visualization software developer. With the strategic partnership, both companies plan to develop new artificial intelligence-powered solutions to provide actionable insights into complex business decisions and grow their respective customer base globally.

Dataiku launches a new version of its unified AI platform, with control center for AI governance

By

Preetipadma K

-

December 3, 2021

This week, Dataiku debuted version 10 of their unified AI platform. Dataiku is an AI platform that allows businesses to turn raw data into advanced analytics and data visualization. It provides a seamless online version of the Enterprise AI platform to quickly and successfully promote development in smaller businesses.

Dataiku 10 comes with a built-in set of tools to assist IT administrators and data scientists in evaluating, monitoring, and comparing models in development or production. With integrated industry solutions, dedicated workspaces for business users, and accelerators for exploratory data analysis, geospatial analytics, and machine vision, Dataiku 10 enables enterprises to provide AI results quicker. Furthermore, Dataiku 10 helps businesses provide value quicker by combining industry-specific solutions, enterprise work environments, faster exploratory data analytics, spatial data analytics, and photo and video analytics (computer vision).

The new Dataiku 10 version has received a substantial upgrade that focuses on three key features and themes:

Scaling analytics and ML activities in a secure manner

The MLOps capabilities in Dataiku 10 have been enhanced to enable IT operators and data scientists to analyze, monitor, and compare machine learning models in development and production. Teams can gain a better understanding of the behavior and performance of live models with automatic drift analysis and enhanced “what-if” simulations.

In an attempt to ease the work of data scientists, Dataiku’s AI platform offers no-code GUIs for data prep and AutoML features to do ETL, train models, and assess their quality. This feature is designed for technically skilled users who do not know how to code and allows them to perform numerous data science activities. Users can use a no-code GUI to decide which ML models the AutoML algorithm can utilize and conduct simple feature manipulations on the supplied data.

Following training, the page offers visuals to help in model interpretability, including not just regression coefficients, hyperparameter selection, and performance indicators, but also more advanced diagnostics such as subpopulation analysis.

Governance and oversight

AI professionals, risk managers, and key stakeholders may use Dataiku’s AI Governance to comprehensively manage projects and models, as well as monitor the overall progress of the AI portfolio. Customers can view all of their models in a central model registry, regardless of whether they were created in Dataiku or with third-party tools like MLflow. Superior AI oversight is provided through structured frameworks for project workflows, approvals, and project qualification.

Accelerate value realization with business solutions and accelerators

Customers can choose and harness the appropriate tools for their goals from projects produced for various use cases in many sectors, and organizations can expedite the pace of value realization. These advanced tools include new geographic analytics, native deep learning capabilities, aided data exploration, and better visual and interactive insights.

According to Clément Stenac, CTO and co-founder of Dataiku, “We have always been convinced that non-specialists must also be involved throughout the organization in order to be able to apply AI on a large scale. Dataiku 10 makes that possible.” By focusing on the above three core areas, Dataiku 10 aims to increase the engagement of AI adjacent roles such as IT operators, risk managers, project managers, and domain experts.

Government to reduce Road Accidents by 2024 using AI

By

Dipayan Mitra

-

December 2, 2021

Minister for Road Transport & Highways in the Government of India, Nitin Gadkari, said that the govt. aims to reduce the total number of road accidents to half by the end of 2024. According to him, road safety is now the government’s top priority, and they plan to use AI solutions to achieve their goals.

Currently, India records 150,000 deaths and over 300,000 serious physical injuries annually because of road accidents. The government will use artificial intelligence to manage and predict traffic conditions.

The AI solution would help authorities drastically reduce any chances of a traffic mishap. Gadkari said, “We have implemented advanced technology and electronic monitoring systems for safe and efficient traffic movement on Indian highways. We have implemented several initiatives in collaboration with governments, NGOs, industries and various institutions within civil society.”

He further added that the government understands the need for collaborative effort at the grassroots level to minimize road accidents. The government plans to deploy artificial intelligence solutions for the following purposes-

Improving lane discipline on National Highways.
Vehicle Overspeed and seat belt detection.
Post-forensic accident investigation.
Accident patterns with black spots.
Driver fatigue and sleep indicator.
Advanced vehicle collision system.

“AI can be used to combine data from all applicable sources above with data used to make appropriate changes at the policy level. The private sector has expanded cooperation and can combine it,” said Gadkari. He also mentioned that the government has plans to provide technical assistance to engineering colleges and institutions to investigate and submit accident analysis reports using artificial intelligence.

Chandigarh authorities also took a similar kind of approach earlier this year to effectively manage the city’s traffic conditions. The artificial intelligence-powered system analyzes video footage collected from various CCTV cameras for calculating the duration of traffic lights to maintain a smooth and uninterrupted flow of traffic.

Amazon launches AWS RoboRunner to build robotic applications

By

Ishwarya M

-

December 2, 2021

At the AWS re:Invent 2021 conference, Amazon launched a new robotic service called AWS IoT RoboRunner. With IoT RoboRunner, developers and enterprises can build fleet management applications for robots.

RoboRunner assists enterprises in managing and optimizing the lifecycle of various robot fleets since it offers an automated infrastructure for fleet management. This service can be used in public warehouses, hospitals, retail shops, supermarkets, shipping harbors, and even in homes for domestic purposes.

Usually, enterprises rely on domestic robots made by different vendors to perform various fleet management operations. However, each robot has its own control software, data format, and processing speeds; it is difficult for enterprises to use robots with a centralized application.

Since each robot has a unique architecture, it becomes difficult for enterprises to build applications for managing robots according to the use cases. To make effective robotics automation systems, companies are required to optimize robots according to their applications and task orchestrations to allow individual robots to perform a series of tasks together.

But, it is challenging for enterprises to build and deploy different robots for different operations; it requires complex integration and high technical knowledge for building advanced software.

To ease the process of deploying different robots in the enterprise’s warehouse, AWS is releasing IoT RoboRunner after finding success within their own warehouses to manage fleets of different robots. Amazon was able to manage more than 350,000 robots in their own fulfillment centers and warehouses worldwide.

To have a preview of AWS RoboRunner, check this link.

Indian Government launches Face Recognition system for Pensioners

By

Dipayan Mitra

-

December 2, 2021

Indian government face recognition pensioners

The government of India recently announced the launch of a new face recognition system that will be used to provide certificates of life. The newly launched system will help pensioners and elderly individuals to quickly receive their pensions.

According to the government, the face recognition system will improve the ease of living for senior citizens. The new system was unveiled by the Minister of State for Personnel Jitendra Singh on Monday. The system will be practical and useful for those senior citizens who are incapable or have problems in submitting fingerprints as biometrics proof.

The Ministry of Personnel, Public Grievances and Pensions mentioned that the face recognition system would affect 68 lakh retired government employees along with workers under the Employee Provident Fund Organization and state governments.

Minister Jitendra Sing said, “the central government has been sensitive to the needs of pensioners and to ensure the ease of living for them. Soon after coming to power in 2014, the government decided to introduce and implement digital certificates for pensioners. This unique technology will further help pensioners.”

He also thanked the Ministry of Electronics and Information Technology and the Unique Identification Authority of India for developing this face recognition technology. The launch marks the introduction to a standard software named Bhavishya, which will be used to process every pension case. The software will be used by all the ministries of the Indian government.

To use this face recognition service, individuals must have a smartphone with a 5-megapixel camera, internet connection, and Adhaar card. According to PTI, “The identity of a pensioner or family pensioner will be determined using a face recognition technique under this capability. Any Android-based smart phone will be able to submit a Life Certificate utilising this technology.”

Interested people can visit here or download the smartphone application named AdhaarFace ID to register for this facial recognition system.

Last few days to apply for Generation Google Scholarship for Women in the Asia Pacific region

By

Dipayan Mitra

-

December 2, 2021

Technology giant Google earlier had launched its Generation Google Scholarship program in the Asia Pacific region. The program is meant exclusively for women and would provide them with training in computer science.

The Generation Google Scholarship program would help students to inculcate critical skills related to computer science for making them industry-ready. The program has been available for quite some time, and the last date of submitting applications is round the corners.

Eligible applications should not miss this opportunity and submit their application forms by 10th December 2021. Google encourages women with a keen interest in computer science to apply for this unique program. The selected applicants will receive a scholarship of $1000 for the 2022-2023 academic year.

Below mentioned are the eligibility criteria to apply for the Generation Google Scholarship program –

Applicants must be enrolled in a full-time bachelor’s degree program for 2021-2022.
After the completion of this scholarship program, applications must be in the second year of their bachelor’s degree from a recognized university in the Asia Pacific region.
Applicants must be enrolled in computer science, computer engineering, or other programs of related fields.
Must have a good academic record.
Applicants should have a passion for improving the representation of underrepresented groups in computer science and technology.

Interested candidates can apply here.

Apart from the Generation Google Scholarship program, Google recently also launched its new scholarship program to train Indian job seekers in digital technologies. The program is offering 100,00 free scholarships to help job seekers gain the necessary skills that the market currently requires. Google has tied up with many industry-leading companies including, TATA, Accenture, Wipro, Tech Mahindra, and many others, to provide employment opportunities to applicants.

Amazon introduces Trn1 chips to speed up the training process of ML models

By

Ishwarya M

-

December 2, 2021

On November 30, 2021, Amazon introduced three new amazon EC2 instances powered by AWS-designed Chips. The newly launched chips help developers/customers improve the performance and energy efficiency of their ML models. In the AWS re:Invent event, Amazon launched three instances called Graviton3, Trn1 and Nitro SSDs. However, Trn1 gained more attention among ML enthusiasts as Trn1 instances will be capable of delivering bandwidth of about 800 gigabytes per second.

This feature from AWS makes it more suitable for large-scale and multi-node distributed training use cases like natural language processing, object detection, recommendation engines, image recognition, etc. The company claims that these processors are also optimized for high-performance computing, media encoding, batch processing, scientific modeling, ad serving, and distributed analytics.

In the traditional cloud ML process, 90% of the cost of ML operations is spent on performing inference about the ML models. To avoid this, in 2019, Amazon came up with a processor called Inferentia. It delivers the best performance and throughput needed for machine learning inference at a lower price than GPU-based instances.

Similar to the inference process, ML training will also be costly since it requires high-performance computing features with parallel processing methods. To simplify the training process of ML models, last year, Amazon introduced a Trainium chip that is specifically designed for machine learning models.

Yesterday, Amazon released the Trn1 chip, considered a sequel of previously launched Inferentia and Trainium chips. The critical feature of the Trn1 chip is that it boosts ML model training by internally performing highly parallel math operations with the highest computing power. The newly released chip provides a 25 percent higher performance compared to previously launched chips.

In Trn1, the company doubled the networking bandwidth to 800 gigabytes per second from 400 gigabytes per second, which is the bandwidth of previous chips. The increase in bandwidth brings down the latency and provides the fastest ML training methodology available in the overall cloud services. The Trn1 instances can be combined with thousands of instances to train even the most complicated machine learning models with trillions of parameters.

To have a preview of Trn1 instances, visit the link.

Nasscom launches a new Center of Excellence for IoT and AI in Visakhapatnam

By

Dipayan Mitra

-

December 1, 2021

The National Association of Software and Companies (NASSCOM) has partnered with the Ministry of Electronics and Information Technology and the government of Andhra Pradesh to launch a new Center of Excellence for IoT and AI in Visakhapatnam.

The center is located in the Andhra University Campus and will promote the development of innovative emerging technologies in various sectors, including robotics. The center will provide all the infrastructure required for developers to build cutting-edge solutions and also help promote entrepreneurship by providing incubation facilities.

The new Center for Excellence was inaugurated by Union Ministers Rajeev Chandrasekhar, Mekapati Goutham Reddy, and various other government officials, including P. V. G. D. Prasad Reddy. Researchers in the center of excellence will have all the necessary facilities to develop solutions for real-world challenges.

President of Nasscom, Debjani Ghosh, said, “The COE’s, the Center of Excellence have become almost a melting point that beautifully connects the different ecosystems to understand the big problems that technology can solve, brainstorms the best use technology to address these challenges or problems, and jointly co-create solutions.”

The newly launched Center for excellence will use the capabilities of artificial intelligence and the internet of things to bring in revolutions for industries, startups, and academia. Minister for Industries & Commerce, Information Technology and Skill Development, Government of Andhra Pradesh, Mekapati Goutham Reddy, said that if they can become world leaders in nine advanced technologies, namely artificial intelligence, robotics process automation, edge computing, quantum computing, virtual & augmented reality, blockchain, IoT, 5G, and cyber security, then the state can reach a trillion-dollar economy.

“It is absolutely essential that the Center of Excellence becomes not just academic extensions of university, but they become living, breathing growing centers of energy, dynamism, entrepreneurship, and technology development of the kind that we must deliver on in the coming months and years,” said Union Minister of State for Skill Development and Entrepreneurship & Electronics and Information Technology, Rajeev Chandrasekhar.

AI-powered Lie Detector tool that reads Micro Facial Expressions

By

Dipayan Mitra

-

December 1, 2021

Scientists from Tel Aviv University have come up with a new AI-powered lie detector system that reads micro facial expressions of humans to determine the authenticity of statements.

Professor Yael Hanein and Dino Levy led the research team during the development of this new artificial intelligence-enabled lie detector. Researchers conducted several experiments where they tracked and analyzed micro-expressions of humans that disappear in 40 to 60 milliseconds.

The technology is yet not perfect, but according to the researchers, it is the most accurate facial recognition lie detector tool that has been developed to date that delivers over a 73% accuracy rate.

Researchers experimented on 48 individuals and asked them to pull eyebrows or cheek muscles while developing the lie detector system. The system uses electrodes that motors muscle movements near eyebrows and cheeks to generate results.

Behavioral neuroscientist and project co-head, Dino Levy, said, “We successfully detected lies in all participants and did it significantly better than untrained human detectors. Interestingly, individuals who were able to successfully deceive their human counterparts were also poorly detected by the machine learning algorithm.”

According to experts, the newly developed artificial intelligence-powered lie detection tool is extremely promising and can play vital roles in various industries, sectors, and security areas, including border security.

“Since this was an initial study, the lie itself was very simple. However, in reality, longer lies have chunks of deception and truth both,” added Levy.

The research team is now training the AI lie detector tool using advanced machine learning techniques based on the data collected from the trials conducted during the initial stage.