Archive for the ‘Cloud’ Category


The Visionary CEO of GTCOM Talks Artificial Intelligence, Machine Learning, Big Data, VR, and More

By Loren Wiener (tech-news.tv)- TNTV
Brisbane QLD, Australia
August 2017

To the uninitiated, Interpreters and Translation services can seem like a bit of a secret art, for business and entertainment purposes only. A website being translated from a foreign language to English, a UN translator interpreting a conversation seems hardly life changing. However, translations in a court case working around court room legalese in a foreign language to the defendant can leave lives in the balance even in Australia where there are 250 Aboriginal languages. In Australia and Globally, up till now much if this has all been done with little use of technology to help, and can take years of training and mastering due to the complexity of it all.  The theme at the 21st and 2017FITCongress, International Federation of Translators (Fédération Internationale des Traducteurs) or FIT, 3-5 August is Disruption and Diversification.

As technology evolves with cloud computing, artificial intelligence, (AI), machine learning, big data, and even in the future more quantum computing and VR use on the horizon, the use of technology has moved at an amazingly fast pace. However, having a way of managing the data and technology to make the best use of all of this requires special skills from a company with the pedigree and a best of breed vision on the way forward.

No better Company is showing this proof than Global Tone Communication Technology Co., LTD – (GTCOM) The man with the vision is CEO Eric Yu. I had the opportunity to visit and chat with Eric Yu to talk about this and much more.

TNTV: Hello Eric nice to meet you and thank you for having the time to talk. How did this all begin for you?

Eric Yu: After my dream to be a translator, I worked with the Beijing Olympics as a translator for the athletes. We then got our first call centre and could accumulate a lot of data that the company has already accumulated over 40 years.

TNTV: How does that impact translation technology.

Eric Yu: – All data is good data but some is “weighted” with different values of importance. Combining this with AI, speech recognition and deep learning technology and open source relationships, allows GTCOM to create a best of breed models and predictions.

Eric then shows me some GTCOM live examples on the YEEKIT  software at their stand on how they can use this technology on what appears simple to view and understand.  

Eric Yu: A large part of our focus is “open platform” information, by using all this we can better predict stock trends for instance because of the amount of data being used but managed better and with more value and better predictions. This platform is named “JoveBird” which is a big data solution for intelligent investment. For example, using all this information, (Eric shows me the stock price of a large company on a live map) Days prior to them having a significant share price drop using such things their open source values like, analyst ratings, business activities, financing actions, partnership activity  and social media (

For example, using all this information, (Eric shows me the stock price of a large company on a live map) Days prior to them having a significant share price drop using such things their open source values like, analyst ratings, business activities, financing actions, partnership activity  and social media (chatter), and even a problem with a popular game, all lead to us being able to predict the share drop very accurately offering not just a view of predicting share price but also share sentiment (how bullish or bearish the stock is) . All this data using “events” like product service, revenues, employment actions, credit ratings, assets etc, and all able to be viewed on the level of relevance, using machine learning, cloud, and big data better and more efficiently using more levels across languages. This allows more accurate analysis not based just on stock trends.

TNTV: Can you tell me a little more about YEE Solutions as a Product?

Eric: YEEKIT has 4 types of accounts Standard, advanced, business and customised all offering different features, from cloud storage to industry specific categories. Predefined or user defined.

TNTV: Can you give me a view of how much data we are talking about and how you’re thinking about the transition from Big Data to AI.

Eric: For instance, I thought, what if we took google just as 1 example. We could capture every google item, and every keyword, not just English by all the major languages not labeling just as a language but all the major languages. The power of this was highlighted when Google announced a move to (neural machine translation) (NMT) moving from SMT (statistical methods) a milestone in machine translation.

TNTV: But it is not about replacing people

Eric: – No it is about giving the people better data able to be used more effectively.

TNTV: – SO congratulations, Eric, GTCOM is a relatively new but very exciting innovative leader in machine translation, big data, and artificial intelligence, can you discuss some of the target clients

Eric Yu: – There are many examples, governments from a security perspective may want to better monitor issues like terrorism, stock predictions, and our customers already include Amazon, Alibaba, and others.

TNTV: How all this work together for a solution as I see GTCOM products are divided into 3 branches, language technological ecology, platform & big data technology.

Eric: The products of GTCOM generally have classified into two categories: YeeCloud & YeeSight. YeeCloud is more likely to focus on language translation, which targets the market of language service. YeeSight provides a comprehensive and cross-language, big data service. YEEKIT is the language technological platform which can be regarded as a computer aided translation tool. It is part of YeeCloud ecology.

TNTV: GTCOM is a leader in Machine Translation. What makes it stand out?

Eric: While some companies will spend billions on building their services, we use many open source technologies, combined with the history of 40 years of data gives a better data, form a more diverse source of data being open.

TNTV: GTCOM was a founding part of the LBDA – Language Big Data Alliance – Can you explain a little about the importance of the LBDA?

Eric: Again, the benefit goes both ways, established by government departments, working on enterprises and public institutions, colleges and universities, and scientific research institutions for promoting the mutual application and sharing of language big data for the mutual application and sharing of language big data and related technological resources in the fields of teaching, research and more.

TNTV: GTCOM is working with many including Australia and New Zealand on MT (machine translation) and AI, can you talk a little about the time frame of this exciting work, and what the expected outcomes are?

Eric: Universities again are benefiting from the Open platform offering new technologies and solutions that would not be seen otherwise. In the case of UNSW, they have advanced project based learning, real clients of their own, and new curriculum, new degrees in translation, and interpreting addressing MT and able to lead in research in Language technologies. This means as we are using open source they also are moving towards VR, AR, Big Data and best practice using YEEKIT, YeeSight, etc.

TNTV: YeeSight and YeeCloud have won multiple awards for innovation, in big data and AI – what do you see the next big advances, even smarter AI?

Eric: Yes, better AI, quantum computers technology is on the horizon, as is VR elements to allow it all to work better with the cloud.

TNTV: How is GTCOM able to maintain this momentum of change to stay ahead, is it in R&D or do you see the need for improvement able to slow down? – For example, we saw this recently with mobile phones and even computers, where the speed of improvements was improving dramatically year after year, (Moore’s Law) but the PC market has flattened, some mobile phone companies are feeling this problem also.

Eric: It is a balance, again due to the open platform things evolving do so from many directions. Sometimes the software is ahead of the platforms used, sometimes it is the other way around, but it all must equal a good user experience.

TNTV: What is next for GTCOM?

Eric: We are currently working towards an IPO soon .

TNTV: Congratulations, very exciting times, thank you for your time Mr. Eric Yu.

Advertisements

Loren W
Melbourne Australia

Is JB Hi-Fi Market Recovery Short Sighted?

I have to admit I read with some interest a recent article (worth a read) by Roger Montgomery, founder of Montgomery Investment, in the weekend Australian. His observation on JB Hi-Fi and the concealment of a more complex reality, though I admit I agree with, I believe missed a few key points, though some were touched on.

Firstly, the US and UK market and the majority of the online market unlike the Australian brick and mortar market, is not built on haggling.  This means that almost anyone walking into JB and making even the smallest of purchases can haggle their way to a lower price. This is starting to show its age. I cannot walk into an Apple store and haggle a price, though they also are seeing the light and offering to price match against the likes of JB Hi-Fi and others, adding to the market competition.

Price haggling in this way as JB and others do is a dying trend but one that falsifies Roberts own perspective on JB being higher than the US market on list price alone. The point you missed is that because the perception in some cases and reality in others and the advent of online business means a haggled price though the norm does not help with perception regardless. Consumers are more market research savvy than ever, and an item that is $400 at JB say a new tablet, might be 20% less online. For many consumers, it is not worth trying to haggle when you can buy even from the US and have an item delivered 10-15% cheaper (even though JB would might have matched it if asked). In regards, to competition Apple recently announced they would price match other retailers on their own products, making the competition even hotter, there. Finally there is the dreaded Moore’s Law impact which observes the speed the market changes due to the speed the underlying technology improves. This means with new technology coming out so fast it makes the previous items outdated faster than ever, that the level of inventory required (and a risk) and resulting impact on profits is not surprising going to take a massive hit.  In the last 3 weeks alone we have seen a new Operating System for Microsoft, and Apple; a new iPhone beater for Samsung, a new office suite for Microsoft, a new tablet by Google all being released and a rumored new iPod, iPad, and iPhone (and Samsung again) in the next 3 weeks.  This trend and speed to market has resulted in many of the JB Hi-Fi’s peers in the US and UK going bump and many others going online only to reduce their product overheads. Pricing is a mindset; when was the last time you walked into Coles and negotiated over a bottle of milk. This duality of looking for a better price online and not knowing for sure what the lowest price is in your own retailer is silly at best, or market suicide, and the times have changed.

I recently had a chat with an area sales rep for Canon Camera Australia, that I cornered when I found the latest camera the 7D, had an RRP of $2100 in Australia, $1700 in the US and $1200 in Hong Kong and a cost to JB for $1600 but I was able to purchase it from an importer across the street from Canon HQ in Victoria for $1300, with the full warranty supported globally by Canon (cheaper if you buy online direct) Previously some folks would not have supported the gray import warranty. Canon had a different view and does not use the ‘Gray import’ label instead opting for a ‘Parallel import’ label. Their view is simple and one that folks like JB have to be worried about. Canon’s view was that though the retailer in Australia might not like it, a market savvy consumer should not be punished for their purchase by not having their warranty supported. It was highlighted that this trend is growing and folks like JB Hi-Fi will not be the only ones impacted.

Instead of price deflation what Australia needs is honest price reality, so the lowest price is always offered. Expect to see many other stores impacted in the same way.

Loren is founder of tech-news.tv and a self-proclaimed ‘technologist evangelist’, introducing new products and services for the likes of British Telecom UK (BT) and NEC Australia Japan, including business broadband for BT in 1994, and Data Centre Service for NEC.  He is a beta tester for Google, Microsoft, and Apple. His twitter followers (over 65,000) follow his views and blogs from all over the world He is contactable on twitter @mr_internet and on email at loren@loren.net.au

When iTunes Match launched, the world celebrated. Apple offering cloud services we could all use.

Uniquely, iTunes Match, lets you use your own music library and Apple, looks at the music and if it already exists in Apples extensive cloud library, it moves none of your music but associates your music in the cloud so you can listen to it on any of you iOS devices like the iPad, Mac, iPhone,  and AppleTV.

This is a better way for those that have broadband (with sometimes slow upload speeds) and especially if you have a large existing music library. Amazon and Google offer similar choices, but require you to upload all your own music. Amazon charges for storage above 5GB.

Unfortunately iTunes Match does not work outside the USA.

Actually that is not correct, it does not work unless you have a US address, & US credit card. The assumption is you must therefore live in the USA.  Apple has not said why.  However there are companies offering these minor details, allowing iTunes Match access for all across the globe.

It is believed Apple did not want to launch the service globally as it does not have the cloud infrastructure in place to handle the traffic this would cause if used in mass.

How it works: Once you log into to your iTunes account and register your US credit card and details, you can pay the $25 per annum to apple.  iTunes uses any outstanding vouchers or credits you have before it uses your credit card. There are only 3 steps after that

As it says you can use iTunes whilst iTunes Match is doing its thing. I have just under 10,000 songs, and this takes up 53GB of space. Apple is limiting space to 25,000 songs, so I find this very generous.

That aside there are downsides, though they are few but the first one is a biggie.

1)      Your Own Corrupt Music Data – It is believed many people will have corrupt data (or meta tags) associated with their music tracks. This means in the traditional garbage in and garbage out world of IT, that Apple might miss associate some of your music library with the wrong songs. This can be dangerous if you over write your master library with that from your new found cloud friends Apple. Note there are some good programs that fix the meta tags either automatically or manually (like Tag and Rename and Tune Up Companion, with the later attaching itself to iTunes so you can drag and drop your music into it to cleanse and correct meta tag issues) . So it is worth backing up your music library so it cannot be over written. If you can wait it is worth cleaning you music tracks to make sure all your meta tags are in order.

2)      Music Quality–  Audiophiles that enjoy the new remastered tracks of everything from the Beatles to the Who, may have ripped their music in raw or high quality that is better than the 255k that Apple is offering (that is overall better than most have with 64k and 128k tracks), but an issue none the same (it might be possible to keep the higher-definition tracks in the cloud I am still testing that now). I estimate 5% of my music is above 255k, or around 25 albums or 500 songs).

Time: This is not a quick process but I hope well worth it. I am currently over 2  hours in and will update this after it is complete.  Step 1, The diagnosis phase took 1 hour, Step 2, the matching phase took an hour also matching about 3 songs a second Competition: The Google Music Service as mentioned is similar but it took 21 days to upload half my music, a daunting task indeed, but the service works very well, also globally, and on any browser (phones, tablets, pcs, or macs)