Search Results
keyword：Christy Lin，
4 result(s)
2017/04/27
The introduction of the deep Learning (DL) approach to machine learning has led to incredible developments in artificial intelligence (AI) over the past few years, attracting interests from cloud computing and semiconductor chip businesses According to TrendForce’s latest research on AI, the three forces that drive advances in machine learning are software, hardware and data AI systems based on DL requires vast amounts of data to train their abilities such as identifying objects and processing requests Therefore quantity and quality of data directly affects AI systems’ accuracy As cloud and software platform companies compete to bring AI solutions for different applications, ownership and access to data will become crucial to their strategies
According to TrendForce’s research, finance and data security together have become the largest application category for machine learning/AI solutions, currently accounting for about 20% of the total value of the machine learning market The second largest application category is digital advertising technologies, which makes up roughly 18% of the total market value Other applications for AI are also growing rapidly
Cloud and software platform companies actively seek data control and access
At present, cloud and software platform providers are major players in the development of AI solutions for different applications These players include Google, Amazon Web Services (AWS), Facebook, IBM, Microsoft, Apple, Baidu, Tencent and Alibaba Cloud and software platform companies have huge databases that encompass Internet users’ online activities and information extracted from the usage of their software tools These databases benefit the development of APIs and SDKs for AI systems
“Owning databases and having access to them sometimes are two separate things,” said TrendForce analyst Christy Lin “Providers of public and private cloud services such as AWS, Microsoft and Google do not necessarily have ownership of data from users and businesses However, these major technology brands can have access to the data when users and organizations use their cloud-based APIs The data can be used for the development of AI-powered products and the optimization of the software tools”
Lin also noted the promotion of open source frameworks is also an important part of the AI strategy Cloud and software platform providers want more developers of AI solutions to use their preferred open source frameworks because this creates follow-up opportunities linked to offerings of software and hardware support
Some examples of these open source frameworks include Google’s TensorFlow, Facebook’s Torch, Microsoft’s CNTK and Intel’s Neon (which the company obtained by acquiring Nervana Systems) The competition in the open source framework market is intensifying due to the arrival of new entrants
Devising AI systems for specific professional fields will require domain knowledge
Domain knowledge – data relevant to certain professions – will be a necessary component in improving the accuracy and reliability of AI systems For instance, an AI system that is be deployed in the healthcare field (ie digital diagnosis assistant) has to be trained on data such as medical images and patient health records
Global market intelligence firm TrendForce has established highly specialized databases that record industry and market trends for over a decade With the arrival of AI-based economy, we will continue to gather critical information from the technology sector and provide value-added services
2017/01/19
Development of AI systems and their applications will continue to revolve around voice and image recognition technologies in 2017 According to market research firm TrendForce, the global market value of voice recognition solutions is projected to grow from US$261 billion in 2016 to US$1598 billion in 2021, with the CAGR in the forecast period at 4364% The main source of growth momentum will come from the widespread adoption of voice-based virtual assistants in various devices
Virtual assistant developers look to different smart products as they strive to establish the dominant voice control platform in the IoT market
“Amazon did not reveal new products at CES 2017, but its virtual assistant Alexa garnered the most attention during the event because it was featured in numerous smart products that were being demonstrated,” said Christy Lin, TrendForce “Amazon has been bullish on the market potential of voice control platforms Alexa’s applications have also expanded rapidly as Amazon took the technology beyond its own Echo speakers to other products and released the virtual assistant’s API and SDK to third parties Amazon ultimately wants to make Alexa the dominant voice control platform for smart products, so that the company in turn will wield significant influence in the development of the entire IoT market”
Compared with virtual assistants from other major technology brands including Apple, Google and Microsoft, Amazon’s Alexa is currently the frontrunner in the market “Apple’s Siri is the earliest market entrant, but the IoT solutions that it provides mainly center on Apple products such as the iPhone Google on the other hand has opened up its voice control platform to third-party developers and released Google Home and Google Assistant to compete against Amazon’s services However, Google’s virtual assistants still lag behind Alexa in terms of hardware-software integration”
“Microsoft, which developed Cortana at a much later time compared with Amazon and Google, is under pressure to catch up in the virtual assistant market,” Lin added “Microsoft also made a significant decision to allow third-party developers to re-use the codes they have compiled for creating Alexa’s skills for building Cortana’s skills This move could affect the competitive landscape in the future”
NVIDIA shines in automotive electronics as the race to develop self-driving cars creates demands for advanced image recognition solutions
The growing interest in self-driving cars has spurred development in the market for image recognition solutions Traditional car brands such as Ford, BMW, Nissan and Hyundai together with new vehicle startups such as Faraday Future used CES 2017 as a platform to demonstrate their autonomous driving technologies The event also saw a wide range of new products from core suppliers of automotive electronics including NXP, NVIDIA, Mobileye and Delphi Upstream hardware suppliers are eager to take advantage of the opportunities that exist in this growing application
Lin noted that the outlook of the self-driving car market is uncertain in terms of applications and demand As for the upstream component market, the outlook and competitive landscape is relatively clearer NVIDIA, for example, has emerged as a leading automotive electronics company Not only NVIDIA supplies hardware for Tesla’s electric vehicles, the chip maker also develops its own platform for autonomous vehicle technology At this year’s CES, NVIDIA’s prototype self-driving car (wittily named “BB8”) revealed that the company has managed to leverage its strength as the leading GPU maker to establish a strong presence in the automotive application It is expected that NVIDIA will introduce a more complete solution package later on
2016/01/25
Facial recognition technology is expected to enjoy accelerated growth over the next five years as its applications emerge in the government, enterprise, finance, consumer and other market sectors According to the global market research firm TrendForce, the market value of facial recognition solutions is projected to reach US$230 million in 2015 and will grow to an estimated US$450 million in 2019 This represents a compound annual growth rate (CAGR) of 1797% in the 2015~2019 period Asia Pacific is going to be the main growth driver of the facial recognition market, and the region presently accounts for almost 60% of the global market
TrendForce analyst Christy Lin said that compared with fingerprints, human faces have biometric features that cannot be easily duplicated Moreover, facial recognition is also likely to be more widely accepted by consumers and has a wider scope of applications While the technology is primarily used in security and monitoring systems, its application is gradually expanding to other areas such as smart retail and mobile payment Last year, Alipay, the online payment platform under Alibaba, jointly launched a payment verification system with Face ++, which is a cloud-based facial recognition platform operated by a Chinese startup known as Megvii Aptly named “Smile to Pay,” this payment verification system lets Alipay users pay for their online shopping simply by taking selfies Other major technology companies that have been investing in facial recognition in recent years and hold related technology patents include Microsoft, Google, Apple and Facebook
Lin also noted that the current mainstream facial recognition systems, which record and match 2D images of individuals’ facial features (ie eyes, nose, mouth and ears), can still be deceived by regular photos and videos However, developers are now advancing towards 3D facial recognition One major type of 3D facial recognition solutions verifies both the image that a face produces under visible light and the depth of facial features with the help of infrared (IR) light Another common 3D facial recognition technology uses IR light dots to measure and verify the distances and depths of facial features The infrared light source that is used to support both types of facial recognition solutions is usually 850nm or 940nm in wavelength
A prominent example of 3D facial recognition technology is the RealSense Camera, a 3D scanning solution jointly developed by Intel and Microsoft RealSense Camera is designed to work with Microsoft’s biometric authentication software known as Windows Hello and uses IR laser to capture facial features for record keeping and verification Since more efficient processors will be needed to manage 3D models stored in the databases of facial recognition systems, Intel stands to benefit greatly from developing and promoting products for this application Biometric identification technology therefore constitutes an opportunity for Intel to turn things around as the chip maker is being hard pressed by competition from ARM and Qualcomm
2015/11/05
The advent of the Internet of Things (IoT) has propelled the demand for better security solutions on electronic devices Apple’s iPhone 5s, which was released in 2013, was the first smartphone to feature a fingerprint recognition system as to further guarantee user privacy Since then, interests on the developments of various biometric identification technologies have been rising Global market research firm TrendForce finds that the market value of biometric identification for financial services applications will increase from US$123 million in 2015 to US$657 million in 2019 This significant growth is mainly driven by innovations within the global finance industry Among regions, Asia Pacific has the largest growth potential followed by Europe
Vein patterns become the favored biometric identifier in the finance industry as security risk rises for fingerprint-based solutions
Fingerprint recognition systems are the most widely available biometric identification technology in the market right now, but fingerprints are easy to duplicate, said Christy Lin, analyst of TrendForce Biometric identification technologies that offer better security, such as 3D facial recognition, iris recognition and vein recognition therefore have become the priority targets of R&D efforts
Vein recognition technology is restricted to checking vein patterns of living body tissues and offers reliable reading Moreover, vein patterns are nearly impossible to counterfeit Many banks worldwide consequently have incorporated this technology into their ATMs to improve the user authentication procedure of these machines
In Japan, for instance, the penetration rate of ATMs with finger vein recognition technology is over 80% Banks there have been quick to adopt better security solutions because they legally and wholly liable for monetary losses due to false user identification made by their ATMs Barclays also brought in this technology last year in September The UK-based banking company will extend this biometric security procedure to its general client accounts if the trial period is successful Bradesco, a major Brazilian bank, offers an even more impressive example as the company has not suffered any ATM-related fraud since it upgraded its machines to identify palm vein patterns in 2007
Lin also noted that banks in China are only beginning to adopt palm vein recognition technology Panzhihua City Commercial Bank, which is based in China’s Sichuan province, has taken the initiative by introducing an ATM that reads the vein pattern of an entire palm The bank’s clients first select palm authentication option on the ATM screen, and then they enter their personal security codes into the machine and have their palms scanned Once the vein patterns of their palms are verified, clients will be able make to deposit, withdrawal and transfer ATM cards are no longer necessary for these tasks
