GOA Accused Many Government Agencies Of Using Unlawful Facial Recognition Technologies

By

-

July 2, 2021

GOA Accused Many Government Agencies Of Using Unlawful Facial Recognition Technologies

The Government Accountability Office (GOA) of the United States recently released a ninety-two-page report, which claimed that multiple federal organizations have been using artificial intelligence-powered facial recognition technology provided by non-governmental organizations like Clearview AI. This research was done under the leadership of the GAO director for homeland security and Justice Gretta Goodwin.

“A goal of this research was to provide the ‘lay of the land’ in terms of federal law enforcement’s use of facial recognition technology,” Goodwin said. But The organization found 13 of 42 government agencies surveyed do not know about all the non-government facial recognition tools their teams use; hence they cannot fully assess accompanying dangers.

Officials said, “By implementing a mechanism to track what non-federal systems are used by employees of the agencies. They will have better visibility into the technologies they rely upon to conduct criminal investigations.”

They also mentioned that by analyzing the risks of using those systems, agencies would be able to reduce privacy threats for the public and themselves. Earlier this year, a more sensitive and confidential report regarding the same issue was sent to the Congress members.

GAO inspected the ownership and use of facial recognition technology by forty-two governmental entities, which confirmed that it owns a facial recognition system. The types of activities they use the technology for, and how they track using systems owned by those outside the government.

It was revealed that many organizations, including the Bureau of Alcohol, Tobacco, Firearms and Explosives, and The Customs and Border Protection, were using non-government platforms Clearview AI and Vigilant Solutions to perform facial recognition searches illegally.

Earlier this year, the Canadian police force was also accused of violating the country’s privacy policy by using Clearview AI’s technology.

Grette Goodwin said that there are numerous ways by which artificial intelligence technologies like facial recognition can be misused. Hence, it is crucial for the government agencies to know about the technology their employees use and understand the risks associated with the use of such unauthorized platforms.

PepsiCo Uses Artificial Intelligence To Bring Products To Life

By

Manisha Kalagara

-

July 1, 2021

PepsiCo introduces artificial intelligence and data analytics to bring each product to life. The AI collects and analyzes the data on potential flavor combinations and product categories, aiding the decision-making process.

Customer research plays a huge role in product release. PepsiCo uses an artificial intelligence-driven tool called Tastewise, which positions algorithms to analyze why and what the consumers are eating. Tastewise says its tool has monitored more than 95 million menu items, 226 billion recipe interactions, and 22.5 billion social posts, among other consumer touchpoints. These massive amounts of data are utilized to analyze future products.

For data-driven social predictions, PepsiCo utilizes another AI tool called Trendscop. It focuses extensively on social conversations around food in general on Twitter, Reddit, blogs, review boards, etc. The tool measures, analyze the volume and growth of product-related data by picking up relevant context from various platforms.

Not only for the product selection, but even after the product is developed, artificial intelligence and machine learning play a key role in sales and marketing. AI is also used in the quality control of targeted advertisements. Not only these, but Pepsico also mentioned that artificial intelligence is utilized in reducing water consumption in agriculture and manufacturing.
Customer relationship management is crucial for any large customer serving enterprise, natural language processing helps much effectively in registering customers’ questions, suggestions, or complaints.

Tesla To Recall 285,000 Cars To Fix Autopilot Issues

By

Manisha Kalagara

-

July 1, 2021

Tesla will be recalling over 285,000 cars sold in China to address the risks associated with the autopilot feature. This decision has been voluntarily taken by Tesla after the investigation conducted by China’s vehicle safety authority, the State Administration for Market Regulation (SAMR).

The recall includes 211,256 locally-produced Model 3, 35,665 imported ones, and 38,599 China-made Model Ys as their autopilot system is too easily activated, leading to accidents. Tesla will be upgrading the autopilot system for the recalled cars for free.

Due to the flaw in Tesla’s autopilot systems, drivers might wrongly switch on or off the active cruise control in Model 3 or Model Y cars in certain situations like while making sharp turns that would accelerate the vehicle leading to collisions, said SAMR. Due to these ongoing issues in China.

The owners need not bring their vehicles to the service center for the update as Tesla pioneered the software updates over the internet to cars. However, the fix is still considered as a recall states SAMR after the defect investigation.

Though Tesla’s shares dropped by 2.1% on Monday, “We believe that this is Tesla China’s subtle way to make concessions with Chinese consumers and government in light of the recent PR crisis,” said Miles Qianli Dong, a research analyst at JL Warren. He also mentioned that this week’s “soft recall” shouldn’t significantly impact the companies’ sales in the future.

Despite the current issues, Tesla can still export the cars made in Shanghai all over Asia or Europe. But the company’s growth has been going through a rough road for a few months with Chinese consumers and authorities due to safety issues. Tesla has to improve its software to stay in good graces in Shanghai.

DataRobot Raised $250 Million On $6 Billion Valuation

By

Dipayan Mitra

-

July 1, 2021

DataRobot Raised $250 Million On $6 Billion Valuation

DataRobot recently announced that it had raised $250 million in a fresh funding round on a valuation of nearly $6 Billion. According to reports, the series G funding round was led by Tiger Global and Altimeter Capital Management, including other participants like T. Rowe Price, BlackRock, New Enterprise Associates, and Silver Lake.

DataRobot is a Boston-based machine learning startup founded in the year 2012 by Jeremy Achin and Tom de Godoy. The startup offers an artificial intelligence-powered platform called ‘intelligence revolution’ that allows users to develop custom artificial intelligence models without the need of writing codes.

This is possible due to the company’s unique algorithm, which analyzes the data a user is looking to process. Then it searches for a compatible neural network and automatically fine-tunes it to generate the output efficiently.

The startup’s artificial intelligence platform provides a very easy to operate user interface allowing drag and drop feature, enabling users from non-technical backgrounds to build machine learning models.

Earlier this year, DataRobot signed a contract with Palantir Technologies for designing agile, unique, and real-time solutions to help solve the most pressing demand forecasting challenges.

The fresh funding comes after the startup raised $270 million in new venture funding in November. DataRobot first hit unicorn status with a valuation of over $1 billion during its Series D round of $100 million in 2018 and then raised $206 million in a Series E round in September 2019.

According to rumors, DataRobot has consistently been named among the companies that are planning to go public in 2021. However, no official statement has been released yet regarding this.

The company, in 2019, hired the former Tableau chief financial officer, Damon Fletcher, as its CFO. The move was made to give the company a fresh perspective as it approaches a public offering.

European Union Seeks For A Ban On Facial Recognition In Public Spaces

By

Dipayan Mitra

-

July 1, 2021

European Union Seeks For A Ban On Facial Recognition In Public Spaces

European Union’s two major data protection watchdogs, the European Data Protection Supervisor (EDPS) and the European Data Protection Board (EDPB), have jointly raised a demand for banning artificial intelligence-powered facial recognition and other biometric identification in public spaces.

The demand calls for a complete ban on facial recognition, gait, fingerprints, voice, DNA, keystrokes, and other biometric or behavioral signals. Recently, the European Commission proposed a harmonized European Union policy for artificial intelligence.

The proposal aims to tackle the abuse of artificial intelligence technologies for privacy protection. The watchdogs’ demand was welcoming of the many artificial intelligence applications in the European Union, but pointed out the necessity of it being removed from the public spaces and for it to not be used to automatically group people based on perceived intrinsic characteristics.

The proposal allows high-risk artificial intelligence applications to be used in fields like migration and law enforcement, but with adequate precaution, such as a fine of 6% of a company’s annual turnover in case of a breach.

Andrea Jelinek, Chairperson of EDPB, and Wojciech Wiewiórowski, Chairperson of EDPS, said, “A general ban on the use of facial recognition in publicly accessible areas is necessary if we want to preserve our freedom and create a human-centric legal framework for artificial intelligence.”

They further added that the proposed policy must also restrict the application of artificial intelligence for social scoring like the one being developed in China, as it is discriminatory and completely against the fundamental values of the European Union.

The European Union has always been very privacy-conscious, and the proposed ban on facial recognition proves that even further. Though EDPS and EDPB have no direct influence on the lawmaking process, they serve as the highest level of advisors for the lawmakers.

Nvidia’s Vid2Vid Cameo Creates Talking Heads Using Artificial Intelligence

By

Manisha Kalagara

-

July 1, 2021

Nvidia announced Vid2Vid Cameo, an artificial intelligence model that uses generative adversarial networks (GANs) to generate realistic talking-heads in video conferencing from a single 2D image. The AI extracts the key points from the first frame in the video as a 2D photo and later uses its unsupervised learning method to collect 3D key points.

Vid2Vid Cameo was demonstrated for the first time in October 2020. It was fabricated for video conferences, and all it requires is only a single picture and a video stream that dictates the person’s animation. The AI model automatically identifies 20 key points to encode the face structure. The GAN on the receiver’s end taps the information from the keypoints and generates a video to impersonate the appearance of the 2D picture into a 3D one.

A GAN is a two-part model that consists of a generator that creates the samples and a discriminator that differentiates between real-world samples and the generated samples to demonstrate impressive feats of media synthesis. According to Nvidia, the high-performance GANs can create realistic portraits of people and objects that don’t exist.

Vid2Vid Cameo develops the 3D talk head in 1/10th of the fraction of bandwidth initially used for video conferencing. Rather than streaming the entire screen of pixels, the model analyzes the facial points of every person on a call and then algorithmically reanimates the face in the video on the receiver’s end.

Nvidia announced that Vid2Vid Cameo would be available on Nvidia Video Codec SDK and Nvidia Maxine SDK as AI Face Codec to achieve exceptional performance that was done using 180,000 high-quality videos as the training dataset.

Dell Launched Artificial Intelligence Powered 4K Webcam In India

By

Dipayan Mitra

-

July 1, 2021

Dell Launched Artificial Intelligence Powered 4K Webcam In India

Dell has recently launched its ultrasharp premium webcam capable of capturing 4k videos. The webcam is preloaded with an artificial intelligence powered feature named ‘Auto-Framing’ to keep the user in focus and in the center of the screen while video conferencing.

The uniquely shaped webcam uses a Sony sensor that has the capability to record astounding videos in dark or low-light conditions. In addition, the webcam also has a noise reduction feature which is very useful in certain scenarios.

The webcam comes with an 8.3-megapixel Sony STARVIS CMOS sensor with the maximum capacity to capture 4k videos at 30 frames per second. In Full-HD (1080p) or HD (720p), the frame rate can be increased to 60fps. It does not have an in-built mic; hence an external microphone will be required while video conferencing.

The company has added a multi-element lens on top to capture additional light. The Dell UltraSharp Webcam also features Digital Overlap HDR functionality that can be used to capture vibrant colors.

Unlike other webcams, the Dell UltraSharp Webcam allows a customizable field of view that can be picked among 65-, 78-, and 90-degree, along with 5x digital zoom. The webcam also comes with a Dell Express Sign-In feature that uses a proximity sensor to detect the presence of a user and automatically logs them out when they step away.

There is also an infrared sensor to enable seamless log-in using Windows Hello. The webcam comes with a USB Type-C that can be used to connect to a PC. The device can be used on both Windows 10 and macOS, said Dell. It has also received certification from Microsoft Teams and Zoom.
Dell UltraSharp Webcam has been priced at ₹18,999. The webcam will be available for purchase worldwide from 29th June 2021.

Wekaio Announced Support Of NVIDIA’S Turbocharged HGX™ Supercomputing Platform

By

Dipayan Mitra

-

July 1, 2021

WekaIO recently announced that it would support NVIDIA’s turbocharged HGX supercomputing platform. The company plans to integrate its Weka FS limitless data platform with NVIDIA’s all new turbocharged supercomputing platform to deliver best in class performance that enterprises demand.

NVIDIA’s HGX platform currently supports three main technologies, which include the NVIDIA NDR 400G InfiniBand networking, NVIDIA A100 80GB GPU, NVIDIA Magnum IO GPUDirect storage software, and NVIDIA NDR 400G InfiniBand networking.

The close collaborative relationship between the two companies has led to this announcement of the major integration to improve performance. NVIDIA official, Don Harris, said, “Advanced artificial intelligence development requires powerful computing, which is why NVIDIA works with innovative solution providers like WekaIO.”

He further added that Weka’s support for their turbocharged platform would enable customers to set up an industry leading artificial intelligence infrastructure quickly.

The new integrated platform will help consumers get a massive performance boost with relatively less overhead, process data faster while minimizing power consumption, and also increase artificial intelligence and machine learning workload more accurately.

Using this new platform, clients will be able to deploy high-performance solutions to an increasing number of industries that require quick access to data both on-premise and cloud.

Co-founder and CEO of WekaIO, Liran Zvibel, said, “NVIDIA is not just a close technology partner to Weka but also one of the strategic investors.” He also mentioned that the two companies are keen to work closely to ensure that WekaIO’s platform supports and incorporates the latest improvements to the NVIDIA HPC platform as they become available. The company has overcome several challenges in this sector by collaborating with many industry leaders, including NVIDIA. They are looking forward to maintaining such relationships to enhance their technology further.

Dvara Partners With IFFCO Tokio To Launch Artificial Intelligence Powered Cattle Digital Identity

By

Dipayan Mitra

-

July 1, 2021

Dvara Partners With IFFCO Tokio To Launch Artificial Intelligence Powered Cattle Digital Identity

Dvara E-Dairy Solutions recently announced its partnership with IFFCO Tokio to launch its new artificial intelligence-powered platform, Surabhi E-tag. The new platform will be capable of digitally tagging cattle to identify them based on muzzle identity.

Pictures of the cattle are collected via the artificial intelligence-based Surabhi mobile application and stored in hi-resolution images as a unique digital identity. The partnership with IFFCO general insurance company was made to assist the digitization process of cattle identity further to speed up the process of their insurance.

Founder and CEO of Dvara E-dairy solutions, Ravi K. A, said, “Lack of scalable, tamper-proof, unique digital identity of cattle is one of the main reasons for moral hazard, resulting in a higher loss ratio for cattle insurers.”

He also mentioned that using advanced artificial intelligence and machine learning technologies, the prime issue of improving the quality of images while capturing the image could be resolved.

One of the biggest challenges faced by cattle insurance companies is the accurate identification of cattle and the extended processing time. The artificial intelligence-powered smartphone application of Dvara E-Dairy captures images with the mobile phone while simultaneously comparing the cattle’s unique digital identity stored in a protected cloud server and retrieves the results in less than a minute.

EVP of IFFCO Tokio General Insurance, Subrata Mondal, said, “We were exploring a dependable cattle identity process that can be stored digitally and can be accessed anytime.” She further added that they are delighted to partner with Dvara E-Diary to develop Surabhi e-Tag to capture muzzle images accurately.

Conventional tagging methods such as polyurethane ear tags can be easily tampered with and are prone to duplication and fraud. Injectable Radio Frequency Identification tags are expensive and require specialized skills. But Surbhi e-tag enables anyone to easily click pictures to tag cattle, which analyzes the discriminative features and stores the information in a tamper-proof, secure cloud-based platform.

WuDao 2.0 Generates The First Virtual Student

By

Manisha Kalagara

-

July 1, 2021

The Beijing Academy of Artificial Intelligence (BAAI) along with the Chinese government has developed the first virtual student named Hua Zhibing with WuDao 2.0. She (virtual student) began her education at Tsinghua University in Beijing, China.

WuDao 2.0 is one of the most prominent language AI models that comes with 1.75 trillion parameters, beating GPT-3 and Google’s size Switch Transformer. “WuDao 2.0 aims to enable ‘machines’ to think like ‘humans’ and achieve cognitive abilities beyond the Turing test,” said Tang Jie, the lead researcher behind Wu Dao 2.0.

The virtual student model will learn faster than the average rate from the virtual student’s processing capabilities. As a result, the model’s learning levels should improve from that of a 6-year-old to a 12-year-old in a year.

WuDao 2.0 is a pre-trained AI model that assists with simulations of conversational speech, understands pictures, writes poems, and even can generate recipes. The language model was trained with a training system similar to Google’s Mixture of Experts called FastMoE, a Fast Mixture-of-Expert (MoE). FastMoE is an open-source system based on Facebook’s open-source framework, PyTorch, along with available accelerators. FastMoE provides the hierarchical interface to generate the flexible model design quickly and adapt to various applications; it also supports large-scale parallel training.

WuDao 2.0 has an added advantage over GPT-3: it can operate in both Chinese and English. This level of robustness was acquired from training with 4.9 terabytes of texts and images that include 1.2 terabytes of Chinese and 1.2 terabytes of English texts, and 2.5 terabytes of Chinese graphic data.

The researchers also say that the next generation Wu Dao model will predict complex models like 3D structures of proteins, similar to that of DeepMind’s AlphaFold.