Google’s TensorFlow Lite Model Maker to Accelerate AI-Adoption in Workflows

venturebeat | April 15, 2020

  • According to a study conducted by Algorithmia, 50% of organizations spend between 8 and 90 days deploying a single machine learning model.

  • Tools like Model Maker could help companies incorporate AI into their workflows faster than before.

  • Google had unveiled TensorFlow Quantum, a machine learning framework for training quantum models, earlier this year.


Google today announcedTensorFlowLiteModel Maker, a tool that adapts state-of-the-art machine learning models to custom data sets using a technique known as transfer learning. It wraps machine learning concepts with an API that enables developers to train models in Google’s TensorFlow AI framework with only a few lines of code, and to deploy those models for on-device AI applications.


Tools like Model Maker could help companies incorporate AI into their workflows faster than before. According to a study conducted by Algorithmia, 50% of organizations spend between 8 and 90 days deploying a single machine learning model, with most blaming the duration on a failure to scale.


Model Maker, which currently only supports image and text classification use cases, works with many of the models in TensorFlow Hub, Google’s library for reusable machine learning modules. (“Modules” in this context refers to self-contained algorithms along with assets that can be used across different AI tasks.) Essentially, Model Maker applies models trained on one task to another related task at varying levels of accuracy, according to several parameters specified at the outset.


READ MORE: GOOGLE OPEN-SOURCES DATA SET TO TRAIN AND BENCHMARK AI SOUND SEPARATION MODELS


Model accuracy can be improved with Model Maker by changing the model architecture, which requires editing one line of code. After the input data specific to an on-device AI is loaded in, Model Maker evaluates the model and exports it as a TensorFlowLite model. (TensorFlowLite is a version of TensorFlow that’s optimized for mobile, embedded, and internet of things devices.)


Models created by TensorFlowLite Model Maker have metadata attached to them, including machine-readable parameters like mean, standard deviation, category label files, and human-readable parameters such as model descriptions and licenses. Google notes that fields like licenses can be critical in deciding whether a model can be used, while other systems can use the machine-readable parameters to generate wrapper code.


In the coming months, Google intends to enhance Model Maker to support more tasks, including object detection and several natural language processing tasks. Specifically, it says it’ll add BERT, a pretraining technique for natural language processing, for applications such as question-and-answer.


The launch of Model Maker follows on the heels of an API — Quantization Aware Training (QAT) — that trains smaller, faster TensorFlow models with the performance benefits of quantization (the process of mapping input values from a large set to output values in a smaller set) while retaining close to their original accuracy. Earlier in the year, Google unveiled TensorFlow Quantum, a machine learning framework for training quantum models, at the TensorFlowDev Summit.


READ MORE: GOOGLE’S APP FOR PIXEL 4 MAPS HUMAN FACES IN 3D USING ARTIFICIAL INTELLIGENCE

Spotlight

Like many businesses, you're probably worried about security, reliability and following industry standards and government regulations


Other News
AI TECH

AI Optimization Technology Company Nota Selected as NVIDIA Inception Premier Member

Nota | January 19, 2022

Nota (CEO Myungsu Chae), an AI optimization technology company, announced that it has been promoted to Premier status in the NVIDIA Inception program. NVIDIA Inception nurtures cutting-edge startups revolutionizing industries with advancements in AI and data science. The free program has 9,000+ members who are given access to the best technical tools, latest resources, and opportunities to connect with investors. As a startup matures, its Inception benefits also evolve to further company growth. Premier members receive increased NVIDIA marketing support, access to Premier-only member events, and a dedicated NVIDIA relationship manager. Selected as an Inception Premier member through the program's rigorous selection process, Nota, based on its AI optimization source technology, provides NetsPresso, a representative solution, an edge-based intelligent transportation system, facial recognition-based access authentication, and low-power driver monitoring solution in the vehicle. Nota plans to advance its AI optimization technology and meet customer needs in the market through close collaboration with NVIDIA. I am proud to have been selected alongside outstanding global AI companies in the first year of cooperation with NVIDIA." Nota CEO Myungsu Chae. In addition, Nota participated in NVIDIA GTC, held in November, and introduced NetsPresso, a hardware-aware AutoML platform. Nota's AI optimization technology, edge-based intelligent transportation system, and low-power driver monitoring solution will also be presented at 2022 Embedded Vision Summit, the premier conference and expo devoted to practical, deployable computer vision and visual AI to be held in California, USA in June.

Read More

AI TECH

Infosenseglobal Announces a No-Code and Free to Use AI Platform

Infosenseglobal | May 09, 2022

Infosenseglobal announced the community release of ML Sense, a first and unique, no code artificial intelligence platform to develop machine learning models in the most nimble and agile manner. The product comes out of the box with more than 40 machine and deep learning models ready to use. To use this product, end users can ingest data from your local desktop or various cloud sources in CSV or industry-standard formats. This is the very first release of the product since it announced its AI division last in February 2021. Infosenseglobal CEO Hitesh Ruparelia believes that his existing ERP customers could leverage this platform via simple predefined data connectors for finance, manufacturing, etc., that they could use for everyday AI/ML application needs. He believes ML Sense is a game-changer product. With its unique intuitive, and easy functionalities, customers can perform rapid experimentation and attain speeds never seen before. ML Sense is a SAAS offering. The 100% free-to-use community version has significant functionalities such as automatic feature selection and engineering, data transformation, scenario builder, hypothesis validation, etc. He further adds that these features would be our competitive advantage compared to existing products and create a niche market position for Infosenseglobal. With ML Sense, the business end-users are empowered to initiate and execute ML programs without the need for technical staff such as engineers and data scientists. Customers will be able to complete the program in days that would have taken weeks earlier. Additionally, they would need no knowledge or prior experience in programming, statistics, or advanced skills to implement and choose which ML models will work for their business scenario. "The good news is that this is just the beginning. ML sense platform is built on top of various open-source mainstream technologies to automate the lifecycle of machine learning applications. Our next goal is to launch this in the marketplace offered by the major cloud providers. The marketplace provides an option to deploy ML Sense in the customers' existing cloud ecosystem. The idea is to provide an alternative solution wherein the customers are not required to bring the data to our platform. This would address the data privacy and compliance-related concerns of our customers. Additionally, many more exciting features and cloud connectors are in the product pipeline soon to be released." Rajesh Hassija, CTO, Infosenseglobal Inc Established in 2006, Infosenseglobal has headquarters in the U.S. and India and offers leading technical business solutions. Backed by deep expertise and experience, Infosenseglobal specializes in creating and delivering bespoke artificial intelligence and machine learning applications for some of the world's leading businesses.

Read More

AI TECH

SparkCognition Delivers Visual AI Capabilities Across Industries With Acquisition of Integration Wizards

SparkCognition | March 28, 2022

SparkCognition, a global leader in artificial intelligence (AI) software solutions perfected for business, is pleased to announce it has signed a definitive agreement to acquire Integration Wizards, a leader in visual AI. Through this acquisition, SparkCognition expands its IP portfolio to include computer vision capabilities, bringing greater value to its industry solutions. The technology leverages new and diverse data sets, including CCTV feeds, drone footage, video from handheld devices, and existing camera infrastructures. The solution can be deployed in hours or days, and helps address critical problems, including safety, security, visual inspection, productivity, and situational awareness. "With advanced visual AI that can recognize complex scenes and activities we further amplify the value we deliver to our customers while leveraging existing infrastructure investments. Integration Wizards has delivered visual AI at scale for innovators and Fortune 500 customers alike, saving lives, safeguarding assets, and driving demonstrable productivity gains." Amir Husain, Founder and CEO of SparkCognition Since its inception in 2014, Integration Wizards' visual AI technology has provided solutions to global organizations including Hindustan Petroleum, Reliance Industries Limited, Heineken, Xerox, Novo Nordisk, and Johnson Controls. Their customers span 16 countries and service over 20,000 users. Key applications include 24/7 monitoring, enhancing safety and security initiatives, improved process automation in factories and facilities, support for autonomous vehicle operations, and environmental and situational awareness. "As part of the digital transformation journey, HPCL is deploying AI based visual analytics across its retail network to improve customer satisfaction and safety. The solution being implemented covers more than 100,000 CCTV cameras, by far the largest implementation of vision analytics in the country", said Shri Ch. Srinivas, CGM – Digital Initiatives, HPCL. "AI has been instrumental at advancing the use of data within our organization and I am excited by the breadth of AI solutions to help us provide cutting edge solutions for improving customer convenience and service, predict future events, and optimize processes." This announcement comes following a year of record growth for SparkCognition, doubling revenue year-over-year. Last year the company also acquired three AI businesses, expanding into key markets including renewables, financial services, and logistics. Additionally, SparkCognition announced the close of its $123 million Series D fundraising round and a unicorn valuation of over $1.4 billion at the start of 2022. "We are thrilled to join SparkCognition, building on their leadership in AI and proven industry solutions. The team is excited to capitalize on their incredible depth of talent, considerable IP portfolio, and deep subject matter expertise, providing our customers with even greater value and a broader AI based portfolio," said Kunal Kislay, CEO and Co-Founder of Integration Wizards. "The breadth of SparkCognition's AI solutions and services, coupled with our extensive experience in deploying Visual AI capabilities to Fortune 500 customers, will be a game changer."  About SparkCognition SparkCognition's award-winning AI solutions allow organizations to predict future outcomes, optimize processes, and prevent cyberattacks. We partner with the world's industry leaders to analyze, optimize, and learn from data, augment human intelligence, drive profitable growth, and achieve operational excellence. Our patented AI, machine learning, and natural language technologies lead the industry in innovation and accelerate digital transformation. Our solutions allow organizations to solve critical challenges—prevent unexpected downtime, maximize asset performance, optimize prices, and ensure worker safety while avoiding zero-day cyberattacks on essential IT and OT infrastructure.

Read More

AI TECH

New Release of the expert.ai Platform Accelerates AI-based Natural Language Solutions Adoption and Time to Value

Expert.ai | April 27, 2022

Expert.ai today announced the new release of its purpose-built platform combining symbolic, human-like comprehension and machine learning to turn language into data, analyze and understand complex documents, accelerate intelligent process automation and improve decision making. By extending core features and adding unique capabilities, such as out of the box knowledge models and connectors, the new platform release increases flexibility, simplifies integration and optimizes data pipelines to augment efficiency across every process that involves natural language (NL). According to Gartner®, "enterprises sit on unexploited unstructured data, with opportunities to extract differentiating insights. Data and analytics technical professionals must uncover such insights by applying natural language technology solutions solutions: intelligent document processing, conversational AI and insight engines." [1] While data has always been integral to businesses, the proliferation of language within the enterprise and the rise of digital transformation have created a new sense of urgency to increase the adoption of artificial intelligence (AI) technologies for natural language to get the most out of all data available to an organization. This has led to the increasing adoption of AI solutions for natural language to get the most value from all data available to an organization. To scale quickly and make AI investments successful, organizations need the tools and capabilities to deliver impactful results, faster. Expert.ai has been pioneering efforts in composite, or hybrid, AI with proven-market expertise and best practices honed from hundreds of successful, real-world implementations across industries, including insurance, financial services and banking, publishing and media, defense and intelligence. Specifically designed for natural language AI, the expert.ai Platform leverages the combination of different AI techniques (machine learning and rule-based symbolic comprehension) with a simple and powerful authoring environment to effectively support the full natural language processing workflow. It is based on the principle that no single natural language AI technique is a fit for every project. According to Gartner®, "The days of singular AI techniques are coming to an end. Software and service providers that cannot provide solutions combining multiple AI techniques (such as machine learning, rule-based systems, optimization techniques, knowledge graphs, natural language technologies) will quickly find themselves at a disadvantage compared with those that can. The introduction of composite AI techniques, even within existing products, will have a profound impact on their capabilities." [2] Features and benefits delivered by the new release of the hybrid AI expert.ai Platform include: 'Smarter from the start' knowledge models deliver NL applications to production faster with higher levels of business accuracy: new Platform release includes access to customizable pre-built rules-based NLP Knowledge Models used to classify text and extract entities, insights and relationships specific to a domain or use case. Knowledge models made available by the new platform release include: finance (commodities, currencies, macro-economics); ESG (environmental social and governance); life sciences; behavioral and emotional traits; PII (personally identifiable information, with redaction or pseudonymization). Simplified deployment processes across multiple environments, including Azure: new release enables users to select MS Azure as a preferred deployment environment. Easier integration, out of the box connectors: unlock more insights through hundreds of connectors via the Boomi Integration Platform and Qlik Connectors that provide quick, easy and secure access language assets in third-party systems and applications. Enhanced natural language operations: provides the ability to include custom Python and Java scripts or third-party services for pre- or post-processing activities in NL workflow orchestrations. "Turning data into a strategic asset is critical to consolidating value creation and remaining competitive. The new release of the expert.ai Platform extends core purpose-built capabilities and introduces innovation to continue providing tangible business outcomes, enabling organizations to easily leverage their language data and accelerate AI deployments and time to value." Luca Scagliarini, chief product officer at expert.ai About expert.ai Expert.ai is a leading company in AI-based natural language software. Organizations in insurance, banking and finance, publishing, media and defense all rely on expert.ai to turn language into data, analyze and understand complex documents, accelerate intelligent process automation and improve decision making. Expert.ai's purpose-built natural language platform pairs simple and powerful tools with a proven hybrid AI approach that combines symbolic and machine learning to solve real-world problems and enhance business operations at speed and scale. With offices in Europe and North America, expert.ai serves global businesses such as AXA XL, Zurich Insurance Group, Generali, The Associated Press, Bloomberg INDG, BNP Paribas, Rabobank, Gannett and EBSCO.

Read More