Page 12
www.us-
tech.com
Supply Chain
Cognitive Computing from the Shop Floor to the Supply Chain
By Marie Cole, Distinguished Engineer — Systems Supply Chain, and Tom Ward, Cognitive Project Leader — Chief Analytics Organization, IBM
ry floor and the supply chain. “Cogni- tive Manufacturing,” as it is called, is an evolutionary step in computer-en- abled production system control that pushes beyond “smart” technologies. According to experts Elizabeth
A
Hoegeman of Cummins, Inc., and J. Rhett Mayor of Georgia Institute of Technology, intelligence and reason- ing are retained by the user, imbuing the manufacturing process with something akin to perception and judgment. This then enables au- tonomous operation of the system, based on a sort of embedded cogni- tive reasoning, reliant only on high- level supervisory control. Cognitive manufacturing sys-
tems perceive changes in the produc- tion process and respond to dynamic fluctuations by adapting. This allows the process to stay within target
rtificial and augmented intelli- gence capabilities are drastical- ly transforming both the facto-
ranges of cost and production rate. In the factories of the future, ro-
bots and humans will increasingly work together. Most of the robots on the shop floor today are specialized and work independently, separated by protective fences. Due to advances in sensor technology and cognitive technologies, such as machine learn- ing, the adoption of collaborative ro- bots, or “cobots,” will certainly in- crease. Examples include KUKA’s LBR iiwa series and ABB’s YuMI. IBM’s work in this area in-
cludes embedding cognition in ro- bots, such as Softbank Pepper, which allows human-robot collaboration through speech, physical touch and visual recognition. In September 2016, the European Factories of the Future Research Association re- leased a statement on the topic: “Intelligent features and predic-
tion-based reactive control strategies within machinery and robots will
radically change their interfacing with human workers in manufactur- ing environments. This will be in such a manner that the human-robot system will be dynamic, will act safe- ly in a shared working space, will fol- low an intuitive cooperation, and will be aware of its work and of its envi- ronment.” The strength of human action is
based on the ability to recognize pat- terns, evaluate previous learning and experience, and to apply probability- based reasoning before taking action. In the same way, cognitive manufac- turing uses a comprehensive, interac- tive body of information to find new patterns and assist the human-ma- chine decision-making process. Cognitive manufacturing uses
APIs (application programming in- terfaces) to augment human decision making. So, today’s artificial intelli- gence (AI) is more of an “augmented intelligence” that assists with deci- sion making, rather than replacing humans entirely.
Cognitive Computing Fundamentally, cognitive com-
puting capabilities in an agile manu- facturing system can help to:
l Extract useful information from
Trust NTE Electronics to be your Electronic Components Supplier
INDUSTRIAL CONSUMER
COMMERCIAL MRO APPLICATIONS
Cross references for over 525,000 industry part numbers
Expanded range of high quality product offerings
Convenient distributor network for instant inventory fulfillment Professional and knowledgeable staff Quantity discounts available
Active/passive/electro-mechanical/ interconnect and optoelectronics
www.nteinc.com
The company behind the NTE and ECG Brands!
both structured and unstructured data (documents, images, audio, video) to understand context and meaning.
l Reason by using established un-
YOUR SOURCE FOR SEMICONDUCTORS, RELAYS, SWITCHES, CAPACITORS, RESISTORS, AND MORE
derstanding to make connections, form hypotheses and to prioritize suggestions.
l Learn continuously by accumulat- l Interact through a natural multi-
ing data and insight through human interactions.
modal human interaction, including freeform interactive dialog, text, speech, gestures, visualization, and collaboration.
Another important factor is that
cognitive systems are not purely pro- grammed. They are instead trained to acquire information through expe- rience and to learn over time. These systems use such technologies as:
Natural Language Processing (NLP). NLP allows a computer to an- alyze and understand human lan- guage. NLP includes speech recogni- tion, where the computer uses a built- in model to transcribe natural human speech into sequences of words and sentences. The meaning of the words is computed in a single task called POS (part-of-speech) tagging. Then, the text-to-speech task converts the tagging format into one that is able to be understood by humans.
Machine Learning. Machine learn-
ing helps to find highly-complex and nonlinear patterns in data of differ- ent types and sources. It transforms raw data into models, which are then used for prediction, detection, classi- fication, regression, and forecasting. Machine learning algorithms learn from the dynamic system and adapt to changing environments automati- cally.
Machine Reasoning. Machine rea- soning refers to automated reasoning that is associated with the human process of thinking. Automated rea- soning enables a computer to under- stand different parts of the reasoning process and to solve problems that are technically complex. Machine reasoning involves mathematical log- ic, artificial intelligence and theoret- ical computer science.
Visual Analytics. Visual analytics are used to display data in com- pelling ways and do not rely only on language to be understood. Color, tone, language, and mathematical figures are all ways by which a cogni- tive system can understand and make itself understood.
Cognitive Computing in the Supply Chain
Supply Chain Risk Insights, pow-
ered by IBM’s Watson, is an applica- tion that allows risk analysts to inves- tigate risk magnitude, compare it to previous events in the region, and to inform supply chain site leaders. Risk Insights has been deployed
as an internal-use cognitive risk mit- igation system at IBM since mid- 2015. Since that time, it has identi- fied hundreds of potential risk events, and dozens that were in prox- imity to manufacturing sites and supplier locations. This system analyzes social me-
dia posts and other data that are as- sociated with particular events and makes them available to the risk an- alyst in real time. The devastation caused by this year’s hurricanes is an example of the benefits of such a sys- tem. It can identify potential risk lo- cations based on weather data, social media and knowledge of the user’s supply chain. Cognitive manufacturing will
certainly become more commonplace. With a foundation of reasoning ma- chines, humans can maintain the highest level of control, without be- ing responsible for the majority of manual or dangerous labor. By lever- aging the massive potential of ma- chine learning and tethering it to au- tomated production processes, we can greatly increase the efficiency of production, while streamlining the supply chain. r
December, 2017
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52 |
Page 53 |
Page 54 |
Page 55 |
Page 56 |
Page 57 |
Page 58 |
Page 59 |
Page 60 |
Page 61 |
Page 62 |
Page 63 |
Page 64 |
Page 65 |
Page 66 |
Page 67 |
Page 68 |
Page 69 |
Page 70 |
Page 71 |
Page 72 |
Page 73 |
Page 74 |
Page 75 |
Page 76 |
Page 77 |
Page 78 |
Page 79 |
Page 80 |
Page 81 |
Page 82 |
Page 83 |
Page 84 |
Page 85 |
Page 86 |
Page 87 |
Page 88 |
Page 89 |
Page 90 |
Page 91 |
Page 92
