High Performance Computing and the Opportunity with Cognitive Technology
2 likes1,016 views
The document discusses high-performance computing (HPC) and the integration of cognitive technology, emphasizing how systems like IBM Watson can enhance decision-making across various domains. It highlights the capabilities of cognitive systems, their applications in healthcare and sectors like oil and gas, and the challenges organizations face in data management and computational efficiency. The presentation also outlines potential opportunities for utilizing cognitive technology in transforming workflows and improving insight generation.
High Performance Computing and the Opportunity with Cognitive Technology
- 1. High Performance Computing and the Opportunity with Cognitive Technology IBM Watson
- 2. Sridhar Sudarsan CTO and Chief Architect IBM Watson Ecosystem Our Expert:
- 3. Agenda: • What is cognitive? • How cognitive technology is being used today • Cognitive in HPC • How you can get started
- 4. Computing Eras A new era of cognitive computing will forever change the way people interact with computing systems. Cognitive systems will help people extend their expertise across any domain of knowledge and make complex decisions involving extraordinary volumes of information.
- 5. What is Cognitive Technology? Understands and converses in natural language Weighs evidence to provide “best” possible outcome Interacts and maintains context Adapts to new domains by reasoning Applies statistical learning techniques as opposed to rule based techniques Learns by experience or via instructions by humans
- 6. Questions, Answers and Natural Language Processing • Watson used deep Natural Language Processing in Jeopardy! to answer questions and beat the Grand Masters of the game – Deep analysis of language – Hypothesis generation – Measuring inferences to candidate answers – Operating on a massive scale – 200 Million pages of literature – Working at incredible speed – responding in ~ 3 seconds
- 7. In healthcare and many other industries, the input (question) is even more complex Evidence Profile for UTI Diagnosis A 58-year-old woman presented to her primary care physician after several days of dizziness, anorexia, dry mouth, increased thirst, and frequent urination. She had also had a fever and reported that food would “get stuck” when she was swallowing. She reported no pain in her abdomen, back, or flank and no cough, shortness of breath, diarrhea, or dysuria. Her family history included oral and bladder cancer in her mother, Graves' disease in two sisters, hemochromatosis in one sister, and idiopathic thrombocytopenic purpura in one sister. Her history was notable for cutaneous lupus, hyperlipidemia, osteoporosis, frequent urinary tract infections, three uncomplicated cesarean sections, a left oophorectomy for a benign cyst, and primary hypothyroidism, which had been diagnosed a year earlier. Her medications were levothyroxine, hydroxychloroquine, pravastatin, and alendronate. A urine dipstick was positive for leukocyte esterase and nitrites. The patient was given a prescription for ciprofloxacin for a urinary tract infection and was advised to drink plenty of fluids. On a follow-up visit with her physician 3 days later, her fever had resolved, but she reported continued weakness and dizziness despite drinking a lot of fluids. She felt better when lying down. Her supine blood pressure was 120/80 mm Hg, and her pulse was 88 beats per minute;; on standing, her systolic blood pressure was 84 mm Hg, and her pulse was 92 beats per minute. A urine specimen obtained at her initial presentation had been cultured and grew more than 100,000 colonies of Escherichia coli, which is sensitive to ciprofloxacin.
- 8. Capabilities of a Cognitive System
- 9. Watson as a Cognitive Platform
- 10. Watson Ecosystem Partners A leader in the deployment of Natural Language Generation technologies in data analytics and information delivery allows you to ask free form questions around the regulations that are relevant per the report on compliance legislation. A company that helps healthcare professionals offers physicians a patient-facing clinical decision-making mobile platform that physicians can use to make contextual queries against vast volumes of medical research. A software solution that spans a number of key workflows and tasks to help find, engage and collaborate with both external innovators and colleagues to better understand global science and technology resources. Cognitive Security Insights (CSI) is a cloud based solution that delivers an incredibly powerful cyber threat assessment and research platform by assessing the threat and providing context-based research and remediation options.
- 11. Ø The world is awash in data – sensor, modeling, social Ø The vast majority of data is collected but unanalyzed Ø Organizations are experiencing data challenges as the focus on insight Oil and Gas •15 PBs of survey data •10+ months to process new survey data once acquired •100x compute needed for deep water imaging •Network and high bandwidth storage intensive •Long term storage needs •Multi-‐TB datasets for visualization •1 TB per field per day, 2 TB per rig per day Genomics and Life Sciences • NGS data doubling every 5 months • Some leading institutions @ >100PB in 1-‐2 years • Projects increasingly National scale (UK, China, UAE, US) • Large cluster, Large SMP needs • High performance file system and scalable storage • Multi-‐stage workflow process for personalized treatment The HPC Industry Opportunities Challenges Ø How do I bring new approaches to computation? Ø How do I gain more insight from the computational work I am running? Examples
- 12. Disruptive Trends in the O&G industry Data recognized as key to drive value. New methods required for new unconventional exploration, seismic technologies and data overload 80,000 sensors In a single modern platform 94% of production in 1,500 oil Fields (65,000+) 1-2 TB/day One major O&G company only 300,000 Publications / year American Petroleum Institute, SPE, AAPG, SGE, etc Drilling companies report collecting 2TB / day Only 5% gets to the shore 95% is not used Data overload and reliance on traditional, non-scalable methods Workforce transition and shortage of skills IoT, data- and computationally intensive technologies Declining giant fields, slow reserve replacement rates Challenges and risks in new E&P frontiers
- 13. Cognitive computing re-defines workflows across the whole O&G value chain New ventures & appraisals Exploration Field development Production and Operations Asset Optimization Bringing together real- time, unstructured data from a myriad of sources to assess geopolitical risks, local and economic trends, past production history for dynamic decision- making environments Transform seismic interpretation workflow by automated recognition of prospects using cognitive visual analytics contextual knowledge of drilling sites Benchmark with respect to large-scale historical data from similar projects, lessons learned from unstructured data, reports from past projects Assimilation of large amount of continuous spatial-temporal data using machine learnt model blending of physical and data- driven models Analytics augmentation leveraging industrial knowledge graphs for classes of equipment, augmented reality to reduce unplanned downtime Portfolio simulations, (static) correlations / clusters Drilling cost/efficiency Equipment utilization well placement Analyze Project Data Time, Cost, Quality, HSE predict based on history Maximize production efficiency, min operation costs, well productivity, outlier warnings Asset monitoring / management, fault identification, failure prediction Traditional analytics New cognitive-enabled end-to-end transformations
- 14. Cognitive Computing for Life Sciences: tranSMART Challenge: Translational medicine requires sequencing data ingress, data analysis, and secondary analysis. It’s not just the processing algorithm. Solution: Infrastructure biased overwhelmingly towards storage. Elastic Storage Server, Spectrum Scale, Mellanox Fabric, and Power Systems S822LC. Results: Clients have runtime expectations of hours for this workflow. IBM delivers data ingress in <4 min; data analysis in ~20 min. Practitioners focus on achieving insight. Planned Cognitive Enhancement: Watson Text Analytics
- 15. What will you do with Watson? Start building for free on the Watson Developer Cloud: ibm.biz/watson_services
- 16. Questions? Sridhar Sudarsan CTO and Chief Architect IBM Watson Ecosystem Sumit Gupta VP and HPC Business Line Executive, OpenPOWER
- 17. Join our Watson Developers Labs at the GTC2016/OpenPOWER Summit on April 7th in San Jose, CA: ibm.biz/WatsonLabGTC2016 Join us at our next event!