Techtonic 2018

3 min readNov 16, 2018

--

Tecthtonic 2018 is a one day conference organized by Samsung SDS to showcase their enterprise solutions and engage with the community of developers.

The conference takes place in Samsung SDS heaquarters in Seoul Jamsil area on Nov 15, 2018. The event is broadcast live on Facebook Samsung SDS page.

Welcome address speech by Dr Won Pyo Hong (Samsung SDS CEO).

Keynote speech by Dr Shim Yoon (Samsung SDS CTO) “Technology innovation — Partner, disrupt and foresee”.

Samsung SDS has also offices in Israel (Tel Aviv), UK (London) and US (San Francisco) and has partnerships with several universities for research and industrial collaborations: Bar-Ilan Univ., University of Haifa (security), Seoul National Univ. (anomaly detection/homomorphic encryption), Yonsei Univ., Hanyang Univ., Sungkyunkwan Univ. KAIST, UNIST (Brightics Academy), Imperial College (image recognition), Cranfield University (IoT Security), UC Berkeley, Rutgers University (anomaly detection), Georgia Tech (network), State University of New York (text recognition).

Xeed Lab is Samsung SDS internal startup incubation and acceleration program.

Samsung SDS is actively participating and presenting in local and international conferences. Through their open innovation program, SDS has made 65 presentations in events worldwide and around 30 in Korea.

Main clients operate in four fields (manufacturing, finance, service and logistics) and Samsung SDS provides AI solutions for data analytics, image processing and speech recognition with several platforms. Brightics Studio is a newly launched visual solution to perform big data analytics easier and faster. SDS envisions a data driven society, where not only data scientists, but also data based citizens will use big data for decision making for various purposes.

SDS is active in providing BASIC (Blockchain, AI, Security, IoT and Cloud computing) enterprise solutions. In big data analytics they presented a predictive analysis service for connected machineries in smart manufacturing sites. With image recognition defect checks and quality improvements have had significant results in improving manufacturing processes of Samsung Group companies. In speech recognition SDS solutions have provided several automated enhancements: Brighty Work platform helped to reduce manual work requiring two hours to perform in only five minutes.

SDS has developed a blockchain platform, which is faster and more scalable than other public, permissionless and open source blockchain such as Hyperledger or Ethereum. It is also developing and researching homomorphic encryption, which together with zero knowledge proof of concept, may help reconcile the inconsistencies between blockchchain features and GDPR principles.

To foster rapid innovation SDS management combines a top down and bottom up approach to review regularly the road map of the company.

Present technologies are moving from predictive/human for AI to prescriptive/AI for AI solutions. SDS is also preparing for other changes: present siloed data sources may transform in intelligent data lakes; privacy and security concerns will increase with wide spread adoption of AI services; cloud virtualization could be replaced by cloud native computing in the short-medium term.

Keynote speech presented by Young Gun Kim (Python Foundation Director) “Changes and implications in the open source development ecosystem”.

Open source movement milestones: the development of internet, usenet, bbs and newsgroups in the 1960s, Richard Stallman launching the GNU project in 1983, Linus Torvalds developing the Linux OS kernel in 1991, the introduction of my SQL in 1994, PHP in 1995, Apache in 1996 and Python language in 2000.

History of open source, “Evolution of blockchain technology”, Deloitte Insights, Nov 2017.

Open source movement timeline: 1980–mid 1990 embryonic period, end 1990–2010 development time and 2010-present starting to change.

Python seminars and conferences are very friendly open community events. Gender equality is more diverse than Stack Overflow events for speakers.

Python started to become a successful language from the 2.0 release (currently the latest version is 3.0). Node, Python and other open source projects refer to themselves as communities of software developers, contributors, enthusiasts and curious users. Python coding is now used to develop applications for academic research, education, music with internal APIs and even film making (Disney has twelve repositories in GitHub). Python conferences are managed through the Pycon international communities located in more than thirty countries. At Pycon events participants must follow a Code of conduct.

Most popular programming language used in GitHub repositories. Python is among the top five with JavaScript, Java, C++ and Go. From GitHub in blockchain interactive visualizations Deloitte Insights.

“I came for the language, but stayed for the community”, Brett Cannon (Python core developer). Python may not be the best, fastest or easiest to use programming language, but it has one of the largest and most collaborative open communities of developers globally.

Brightics AI is an enterprise web application developed by Samsung SDS to “transform large volumes of data into
easy-to-understand visuals using artificial intelligence” as explained in their website.

The most used programming languages to develop Machine Learning software are Python, RapidMiner, R and SQL. For Deep Learning tools Tensorflow and Keras are the most used.

Brightics Studio is a portable application downloadable on notebooks.

Track 1 “Blockchain/IPA (Intelligent Process Automation)” session video on Samsung SDS Facebook page.

Ki Woon Sung (Samsung SDS Blockchain Research team) “Exploring blockchain platform”.

At the end of 2017 Samsung SDS was involved as the IT partner in an export customs clearance project with other thirty eight companies to develop a marine logistics blockchain. Considering the relatively long navigation time of container cargo vessels, the project required seven months to complete, but only seven weeks to actually program the code. The main challenges were reconciling the needs and requirements of the different stakeholders involved (shipping lines, banks, insurance firms) and coping with the limited pool of blockchain experts and software engineers to rely on. Blockchain can be beneficial for global trade in many respects: 1) providing more end to end visibility, standardization and transparency for data/information/money flows; 2) enhancing trustworthiness with new counterparts, tracking real time the whole processes; 3) reducing transaction costs and other frictions; 4) automating part of the clauses.

In terms of infrastructure, technology and business models fabric (i.e. raw/plain vanilla) and enterprise blockchains differ in many ways: fabric blockchains are open to any kind of contributions (individuals, groups or organizations), enterprise blockchains are built by ecosystems of companies and governmental entities only. Open fabric blockchains are based on decentralized and/or distributed networks, protocols and applications such as bitcoin, ethereum, quantum and EOS, while enterprise blockchains rely on Hyperledger Fabric (IBM), Quorum (JP Morgan Chase), Corda (R3) platforms. In fabric blockchains digital tokens or coins are necessary to reward the effort and participation of stakeholders, dApps are usually developed to increase the utility function of the tokens/coins and funding is done mostly through ICO (Initial Coin Offering); enterprise blockchains in finance and logistics are based on application business models and therefore consortium of companies share the financial and developmental burdens.

“Building a blockchain for business with the Hyperledger Project” YouTube video by IBM developer

Doing business with an open source software and platform has many issues from speed to scalability. Samsung SDS has developed a proprietary consensus model with higher transaction speed, running at 8,200 TPS in Q3 2018, but that can reach higher velocity at 13,000 TPS.

Dataflow uses an acceleration processing, which resembles concurrent transmissions an in old telecommunication systems.

The system interoperability is guaranteed by a multichannel routing solution to exchange data from one blockchain node to another one.

Gu Hyon Park (Samsung SDS Solution Development) “Business automation from bot to chatbot”. Business automation bots are now becoming a real alternative for office automation in paralegal, financial backoffice and for general IPA (Intelligent Process Automation) tasks. Products like Amazon Alexa and Naver Clova are already in the market. Chatbots have the advantage of working 24/7 non stop. However the development behind these automation solutions are laborious: e.g. to automatize data legacy migration work requires six months of preparatory work to program the processes and procedures, while software and script maintenance is also costly. The assessment and definition of scenarios require always human expert supervision and support. Samsung SDS has developed an automated cargo container tracking system, that can check the inside/outside temperatures of specific containers during the entire navigation time on shipping vessels. In smart manufacturing SDS has also developed the concept of LoT (Line of Things) services to complement MES (Manufacturing Execution System) and MRP (Manufacturing Resource Planning) solutions, reducing significant the thousands of manual reports currently required in production or assembly lines. Chatbots and Cobots (Collaborative Robots) can complement and improve humans in many tasks but cannot substitute them.

Starcraft videogame battle between bots during the lunch break.

Track 2 “AI/Analytics”, video replay of the session on Samsung SDS Facebook page.

Track 3 “IoT/Security/SW Engineering”, video replay of the session on Samsung SDS Facebook page.

AlphaLaw is a legal AI solution which can review new legal agreement with standard contracts, assess legal risks and store/order legal documents. Founded by a group of Samsung SDS engineers, who won third prize in an internal hackaton, funded and incubated by Xeed Lab, Alphalaw will be spun-off soon as one of first graduate projects from SDS accelerating programme.

Insfiler is another project funded and incubated by Xeed Lab about AI data profiling solution. Insfiler can produce automatically reports about specific demographics (by gender, age and location) with user defined keywords and with data, statistics and information from forty public databases. The startup is a recent spin-off of Xeed Lab and will be launching the services by the end of the year.

인스파일러:데이터 포탈

인스파일러는 사람에 대한 정보가 필요한 모든 분들을 위한 데이터 서비스 입니다. 인스파일러는 성별, 연령, 지역으로 구분된 가상 인물들의 인구 통계부터 신체 사이즈, 생활 패턴, 소비 패턴을 알수 있는 '프로파일…

www.insfiler.com

Parallel 38°

Written by Parallel 38°

Stories from Korea. In pursuit of truth, impartial pov, fear/favorless, independent of party/sect/interests involved.

Help
Status
About
Careers
Press
Blog
Privacy
Terms
Text to speech
Teams