The latest AI-based model from Fujitsu can reduce the cases of products and services recalling for the company.

Companies across the globe are exposed to a variety of risks. While some of them can be identified and avoided through strategic planning, others can not be even tracked. One of these dangers is a product recall, which normally occurs after a product or a service has been released, thereby adding huge costs to the company and irreversible damages for many.

Fujitsu, a Japanese firm, recently developed an AI system capable of highlighting irregularity in the product’s appearance to detect associated issues at an earlier stage, thereby providing the chance to correct them before the product is released in the market. The AI technology will be used for image inspection, which will allow for the extremely detailed identification of a wide range of external abnormalities on manufactured objects, such as scratches and production errors. 

How does it work?

The particular AI-enabled model is pre-trained on images of the products with simulated abnormalities. The company uses real images of defective goods pulled from a production line’s inspection process for the training data. 

Although many products have similar shape and appearance, the AI-based tool has the capability to correctly identify abnormalities associated with the product. For example — the frayed out threads of the carpet made of different materials or colour or defective wiring patterns on circuit boards can be identified by the AI tool with precision. The Fujitsu lab further confirmed the effectiveness of the AI-model in reducing the man-hours required to inspect the printed circuit-boards by at least 25%.

The earlier methods of training the AI model were based on the tendency to focus on individual characteristics of a product, rather than working on all characteristics of even similar looking products, to identify abnormalities with accuracy. As a result, it is essential to capture a wide range of features of a standard image while training AI to perform quality control tasks. Moreover, it will reduce the workload of the manufacturing industries and enhance productivity.

The Necessity

The first and foremost reason to have more AI-based models is to reduce the enormous cost associated with recalling goods and services. Take, for example, the most recent case of the Hyundai’s battery fiasco. Hyundai had to recall more than 82,000 vehicles – thereby costing the company around $900 million, amounting to $11,000 per vehicle. Similarly, General Motors had recalled around 7 million vehicles due to faulty airbags that hit the company with a whopping $1.2 billion.

Secondly, it creates an unnecessary burden on the companies’ working staff, leading to increased man-hours, overburden of the work, and delays in meeting the targets set by the organisation. This delay is avoidable by emphasising the pre-production phase and adopting AI-based tools for precise product identification. Thirdly, defeated products in the market can cause injuries and fatalities, creating a massive brand image declination. 

Lastly, the Consumers Protection Laws of the respective countries will hold companies accountable for the defects and the harm caused to the consumers. This has been seen recently in March 2021, where Johnson & Johnson (J&J) has appealed with the US Supreme Court in a final effort to reverse one of the country’s biggest product liability verdicts.

The Way Forward

It’s better to embrace the latest AI, ML-based models, and technologies to provide a new life to companies’ production facilities to enhance the final products before rolling out in the market. Rather than facing trials, managing brand crises, or paying hefty sums, companies can look out for deep tech solving the problems and providing a cushion for the long-term good.

The post How Fujitsu Is Using Artificial Intelligence appeared first on Artificial Intelligence.

AI-based detection systems usually work with uncertainty, they are useful not only to raise alerts when something seems to be wrong but also to give a score on how close a given event is from a cyberattack.

Artificial Intelligence (AI) is the ability of technology to think and act like a human and is being utilized to help fight off potential cyberattacks. Employing AI and machine learning to detect vulnerabilities significantly enhances human capabilities, as AI can analyze and report millions of cyberthreats in the time it might take a person to do the same.

AI can get “smarter” and “learn;” as an AI algorithm continues to search and monitor data, it can improve its understanding of diverse types of potential attacks.

According to a study by PwC and Data Security Council titled “Cyber Security India Market: What lies beneath”, Artificial intelligence (AI) and machine learning (ML) will be powering the ‘cyber war rooms’ in organisations to help them protect from increasing cyberattacks, as well as detect, predict and respond to the same.

Defining an attack

You have heard a lot about cyberattacks and data breaches on the news lately, but what exactly is a cyberattack and why is it important to protect our data?

“In short, a cyberattack is a deliberate and often targeted attempt to mount an action via or against digital technology,” says IEEE Senior Member Steven Furnell. “In practice, cyberattacks can take many forms, but we tend to most commonly associate things like hacking and malware, and perhaps phishing. However, there is a far broader range of labels that can be used to reflect the specific approaches and motivations that may be involved.”

“After the attack, the computer or the network will be destroyed, exposed, modified or lose some functions,” adds IEEE Senior Member Guangjie Han. “And your personal information or important data will be acquired by someone else without authorization. Consider the potential benefit behind the data — that’s why these attacks exist.”

Junior hackers might only have an interest in stealing money, acquiring personal data or merely for the joy of hurting others. More senior hackers have the potential to breach entire companies or governments.

How AI prioritizes security needs

Once AI has effectively identified the potential risks and threats, the next step is to prioritize what gets addressed and in what order.

“Since AI-based detection systems usually work with uncertainty, they are useful not only to raise alerts when something seems to be wrong but also to give a score on how close a given event is from a cyberattack,” says IEEE Member Marcos Simplicio.

“Events with higher scores can then be prioritized accordingly. For example, above a certain score, automated measures may be taken to stop the highly probable attack without any human intervention. Events with a lower score can then be forwarded to administrators for further analysis, and some events with low scores may be simply logged without any additional action, since handling them may not be worth the effort.”

Other methods of stopping cyberattacks

AI and machine learning are not the only mechanisms used to prevent and fight off cyberattacks. A wide range of robust technologies protect our systems and help fend off severe breaches.

IEEE Senior Member Aiyappan Pillai explains that in addition to AI, cryptography, emerging quantum cryptography, analytics, IoT security, blockchain security and hardware authentication technologies are also helping technologists keep our digital infrastructure secure.

Lastly, another way to prevent cyberattacks is continuing to educate users to stay away from suspicious activity. “It is important to remember that security is largely a matter related to processes and behaviors and therefore administrative, education and awareness processes play a key role in any well-designed security strategy,” stresses IEEE Senior Member Raul Coucher.

The post How Artificial Intelligence is helping fend off cyberattacks appeared first on Artificial Intelligence.

As more and more people work from home and at times have no control over jarring sounds in the background, Microsoft is set to release a new Artificial Intelligence (AI)-based noise suppression tool in its video conferencing app Teams.

The real time noise suppression in Microsoft Teams will automatically remove background noise during the video meetings.

AI-based noise suppression works by analyzing an individual’s audio feed and using specially trained deep neural networks to filter out the noise and retain only the speech signal, reports MSPoweruser.

“Users will now have control over how much noise suppression they want. The ‘High’ setting is new and will suppress more background noise,” the report mentioned.

The feature is likely to be generally available in November for Windows 10, 7 and 8.1 users.

Microsoft last month announced several new capabilities to help people stay connected, collaborate and build solutions in its video meet app.

Breakout rooms will allow meeting organisers to split up participants into smaller groups to facilitate brainstorming sessions or workgroup discussions.

Coming later this year, custom layouts will allow for a more dynamic content viewing experience and enable presenters to customise how content shows up for participants during a meeting.

Microsoft Teams will also have the option for a recap with the meeting recording, transcript, chat, shared files and more that can be automatically shared in the meeting Chat tab and viewable in the Details tab for each meeting.

Teams will have a streamlined view that shows contacts, voicemail and calling history at once, making it easier to initiate or return a call with a single click.

A new search experience in Teams, powered by Microsoft Search and available late this year, will make finding messages, people, answers, and files faster and more intuitive.

Also, new personal well-being features like a virtual commute, integration with apps like Headspace, and an emotional check-in experience will be available in the first half of 2021.

The post AI-based real time noise suppression coming in Microsoft Teams appeared first on Artificial Intelligence.

Overview of Image and Video Compression

Image and video compression enables you to deliver high-quality media with lower storage and bandwidth requirements. It reduces the size of files, making media easier to transfer and cheaper to store.

To understand how this works, consider any image. Within the image there are many pixels that display the same information, creating spatial redundancies. Compression techniques reduce these redundancies by eliminating or modifying image information. For example, Huffman coding uses entropy coding methods while Discrete Cosine Transform (DCT) uses cosine functions to approximate signal frequencies.

There are a variety of compression standards you can use, depending on the desired outcome. The standard you choose can apply a method such as DCT to identify areas that overlap and eliminate extra data. When this occurs, the quality of media decreases but typically not in a way that viewers can perceive.

This is because high frequencies, such as those created by intense changes or small, sharp details cannot be distinguished by the human eye. The result is that high-resolution images or color spectrums can be reduced to save space.

You can also use methods like DCT on video, however, there are additional methods that you should use as well. For example, interframe compression, which reuses information from initial or subsequent frames to predict frame information. Predictable information can then be eliminated, saving space. Standards that use this method include MPEG and H.264.

Previously, compression was performed only with static algorithms. Recently, however, researchers have developed dynamic compression methods which take advantage of convolutional neural networks (CNNs). These methods use CNNs to perform feature extraction, which can then guide compression algorithms.

What Are Codecs?

Codecs are software or hardware tools containing compression standards. You use codecs to encode and decode data. Codecs are required for compression as well as media viewing or playback.

Some common examples of video codecs are VP9, H.264, and RV40. These codecs are used to modify video streams only, however. To fully compress a video, you also need to use audio codecs, such as MP3, FLAC, or Fraunhofer FDK AAC. In combination, these codecs enable you to compress your videos and all related data, such as audio or title tracks.

One important note about codecs is that a codec is different from a container. Containers are packages that encapsulate all files associated with a video, including video and audio streams, title tracks, metadata, and codecs. Containers are used to interface decoded video data with client players. A container does not dictate how videos are encoded or decoded outside of the codecs it contains. Codecs and containers are frequently confused with each other, because these tools sometimes share the same name. For example, FLAC.

Machine Learning Algorithms for Video CompressionAs AI and machine learning (ML) technologies have advanced, these tools have become useful for improving compression methods. There are multiple ML variations that you can use for compression, but the three main algorithm types are:

• Supervised: Uses predictive capabilities to process and compare large amounts of data in a timely manner. This makes this method well-suited to video processing, making it the most commonly used for encoding and compression.
• Unsupervised: Uses comparison methods to identify similarities in video content. You can use this method to leverage bandwidth economies.
• Reinforcement: Uses feedback loops to refine the algorithm with each successive pass. When these algorithms are used, each adjustment depends on the effects of the previous modification.

The inclusion of machine learning in video compression has the potential to exponentially accelerate the pace of improvement. It can reduce the time needed for compression and the cost of both standard development and processing resources needed.

Machine learning also creates a potential for custom compression algorithms, designed to match each video being processed. This has significant implications for the future of digital media and can help media providers optimize video delivery as they move to the cloud.

The Benefits of Machine Learning and AI for Video Compression

Creating and refining compression tools takes a significant amount of expertise, effort, and time. Integrating AI and machine learning in this process can substantially ease and speed both creation and compression processes. The incorporation of machine learning can provide additional benefits, including:

Faster time-to-market

Machine learning capabilities can enable you to automate compression and codec creation. Using reinforcement learning methods, you can continuously refine both codec development and compression processes with minimal continuing effort. This supported automation helps you optimize your time and speeds overall video processing and delivery.

Improved encoder density

The integration of machine learning enables you to create and use more efficient codecs. Additionally, many machine learning algorithms can be performed in Graphical Processing Units (GPUs) rather than Computer Processing Units (CPUs). GPUs are specialized processors that you can use to perform co-processing in parallel. The use of GPUs can decrease your processing time and improve productivity since CPU power is reserved for other tasks.

Wrapping Up

Current compression algorithms are only able to process a limited amount of data and cannot perfectly correlate patterns. Codecs that incorporate machine learning enable you to process video data with greater detail and accuracy. These tools may even enable you to process data on a pixel-by-pixel level as opposed to frame-by-frame. The result is smaller file sizes with higher quality streams.

The post An Introduction to AI-Based Video & Image Compression appeared first on Artificial Intelligence.

Barcelona-based startup Prometeo has developed an AI-based cognitive health monitoring platform in hope that one day it will be used to monitor the health of firefighters while they are out battling brutal wildfires.

Co-founder Salome Valero explained the development of the platform came about following concerns that were raised by one of their own team members who is a veteran firefighter.

“When the firefighters are fighting against fire, they are breathing in a lot of toxins which can be dangerous for their health … and some of the immediate effects they feel from breathing in smoke is getting headaches. But there is evidence they can suffer respiratory diseases, cancer, and stress disorders,” she told ZDNet, during IBM Cloud Innovation Exchange in Sydney last week.

“The problem is there currently isn’t a lot of data about firefighters’ vitals, but because we have this real-time capability, we are will be able to monitor them.”

Working primarily with firefighters from Catalonia, Prometeo, which was recently announced as the winner of IBM’s Call for Code 2019 competition, has developed a smartphone-sized device that straps on a firefighter’s arm.

It features sensors that measure temperature, humidity levels, and smoke concentration, and can alert in real-time if a firefighter’s health is at risk through colour coded signals of green, yellow, and red.

“We can use it to analyse if the situation of the firefighters is okay, or getting worse, or if it’s a red alert that indicates the firefighter should get out of the fire immediately because their health is in danger,” Prometeo co-founder Josep Rafols said.

See also: IBM announces the finalists of Call for Code 2019 (TechRepublic)

Rafols said even though a minimal viable product has been developed, Prometeo is now looking to improve the product, pointing out how the platform currently relies on Wi-Fi connectivity to transmit data from the senor, which may not be ideal in certain situations.

“Right now, our connectivity solution is using a mobile phone [to hotspot]. We’re aware that some places will not have GSM communication, so one of the things we need to start working on is thinking a little outside of the box of how we will transmit the data when GSM is unavailable,” he said.

At the same time, Rafols said the team recognises the current design needs to be ruggedised.

“The sensors are currently on an arm band that you can just buy from Foot Locker, and obviously that should not be the final solution for the firefighter,” he said.

“In fact, in the initial test, we concluded that having the sensor on the arm is a bad idea because when firefighters clash into a tree the sensor breaks, so we are thinking of a new casing, and the firefighters have said placing it on the chest might be the best approach.”

Eventually, Prometeo hopes through the platform, enough data will be collected in the next two to three years to be used as historical scientific evidence to track the effects smoke has on firefighters’ health.

“This will not prevent fires, this will not prevent natural disaster, obviously, but this will help prevent firefighters from the effects [of firefighting],” Valero said. 

The post Prometeo develops artificial intelligence platform to monitor firefighters’ health appeared first on Artificial Intelligence.