Tech natives 22042013_bartde_witte_watson_v01

471 views
411 views

Published on

IBM Watson

Published in: Business, Technology
0 Comments
5 Likes
Statistics
Notes
  • Be the first to comment

No Downloads
Views
Total views
471
On SlideShare
0
From Embeds
0
Number of Embeds
0
Actions
Shares
0
Downloads
15
Comments
0
Likes
5
Embeds 0
No embeds

No notes for slide
  • Main point: Data is growing at an astounding rate. It is growing so fast that we often lack the ability to use it to its full potential. The highly unstructured nature of this data makes the challenge that much more difficult. This is a real problem for business. It makes informed decisions more difficult to make. Business leaders need a way to find hidden patterns and isolate the valuable nuggets that they need to make business decisions. Further speaking points: Yet, the rewards for finding a way to harness the data into useful information are great; 54% of companies in this year ’s study with MIT/Sloan are using analytics for competitive advantage… and that number has surged 57% in just the past 12 months. “Dying of thirst in an ocean of data”… It’s an apt analogy. Data is everywhere. 90% of it didn't exist just two years ago. The vast majority of it is totally useless for any given goal and therefore amounts to noise and a hindrance to finding the key useful information needed in a specific time and place. Additional information : See information and stats
  • Human performance is one of the things that makes the Jeopardy! Challenge so compelling. The best humans are very, very good at this task. In this chart, each dot corresponds to actual historical Jeopardy! games and represents the performance of the winner of those games. We refer to this cluster of dots as the “ Winners Cloud ” .   For each dot, the X-axis, along the bottom of the graph, represents the percentage of questions in a game that the winning player got a chance to ANSWER. These were the questions he or she was confident enough and fast enough to ring-in or buzz in for FIRST . The Y-axis, going up along the left of the graph, represents the winning player ’ s PRECISION – that is, the percentage of those questions answered the player got RIGHT. CO Remember, if a player gets a question wrong then they lose the $ value of the clue and their competitors still get a chance to answer or rebound . But what we humans, tend to do really, really well is – confidently know what we know – computing an accurate confidence turns out to be key ability for winning at Jeopardy!   Looking at the center of the green cloud, what you see is that, on average, WINNERS are confident enough and fast enough to answer nearly 50% of the questions in a game and do somewhere between 85% and 95% precision on those questions. That is, they get 85-95% of the ones they answer RIGHT.   The red dots represents Ken Jennings's performance. Ken won 74 consecutive games against qualified players. He was confident and fast enough to acquire 60% and even up to 80% of a game ’ s questions from his competitors and still do 85% and 95% precision on average. Good Jeopardy! players are remarkable in their breadth , precision , confidence and speed .  
  • Massive – all possible peaces of evidence Algorithms that analysis evidence Compute confidence scores an rank answers and confidence scores Learn models – machine learning
  • What is bringing about the need for a new era of computing. In large part it is because of the explosion of data. And not just the typical structured data we find in computer databases, but through voice, social media, and sensors throughout the world. Up to 80 percent of this data is projected to be unstructured data by 2015. As you can see, data is just beginning its rapid growth. We’re sill on the blade part of the hockey stick.
  • Healthcare is a great example of how these challenges come to life. Physicians can not keep up with the explosive growth of medical information which is doubling every five years. Reading journals is the primary way new medical information is delivered yet the vast majority of physicians don ’t spend anywhere near enough time to keep up with it.. Meanwhile, diagnosis, treatments, and preventable deaths leave huge room for improvement. Imagine you ’re in a hospital waiting room with 9 others waiting to seen. Chances are, two of you are going to be misdiagnosed. Now sometimes that is not misdiagnosis but delayed diagnosis because the right questions were not asked and the wrong pathway is chosen. Preventable medical errors kill 44K-98K Americans every year. Imagine what the total would be world wide. There is a gap between today’s IT needs and traditional IT methods. Watson is a new approach to IT which can help address some of the healthcare difficulties described here. We are entering a new era of computing
  • • A great way to understand how Watson works is through a simulation. In this hypothetical example, we can see how an incoming patient is diagnosed with increasing precision as more information is made available. <click> • A woman describes her symptoms to her healthcare professional. Based on the symptoms alone, Watson considers a number of possible diagnosis and scores probabilities for each based on the evidence <click> • It then considers explicitly absent symptoms and adjusts diagnosis probabilities accordingly <click> • The iterative process continues, taking into account family history, <click> • … the patient's own medical history, <click> • … medications she is currently taking <click> • … and finally, the results of a simple, inexpensive test to decide between two possible diagnosis
  • Together Memorial Sloan-Kettering Cancer Center and IBM will develop solutions for oncology diagnosis and treatment built on IBM Watson that incorporate the clinical expertise of MSKCC’s cancer experts, patient data, as well as an extensive library of published literature and research on cancer care. This will enable cancer-treating physicians to access relevant cancer care information more quickly in order to personalize diagnosis and treatment plans for their patients. The intent is to make this resource widely available to cancer caregivers so that regardless of where patients live or physicians practice, they can have access to a comprehensive source of information to help them improve patient care. The work that IBM and Memorial Sloan-Kettering Kettering Cancer Center are doing together is related to existing cancer treatment information. This project is not focused on cancer research or finding a cure for cancer.
  • Main point: Join the conversation and take the next step. Further speaking points: . Get involved and learn more about ways that Watson can help your business today. Learn more on the web. Join the conversation on twitter and facebook. See how Watson was created and is having a real impact on youtube. And above all, contact your IBM representative to your priorities and goals and how Watson can help play a part in meeting them.
  • Tech natives 22042013_bartde_witte_watson_v01

    1. 1. Watson from Jeopardy to Healthcare BDent, Bart de Witte, MAppSc Healthcare Industry Leader CEE / ALPS April 2013 – Tech Natives Event, Wirtschaftskammer, WienFollow us @IBMWatsonFollow me @swisshealth20 © 2013 International Business Machines Corporation
    2. 2. watson - jeopardy healthcare & datawatson in healthcare © 2013 International Business Machines Corporation
    3. 3. Jeopardy  Broad/Open Domain  Complex Language  High Precision  Accurate Confidence  High SpeedHuman Language  Words by themselves have no meaning  Only grounded in human cognition  Words navigate, align and communicate an infinite space of intended meaning  Computers can not ground words to human experiences to derive meaning © 2013 International Business Machines Corporation
    4. 4. Why Jeopardy? Grand Challenge © 2013 International Business Machines Corporation
    5. 5. The world is “dying of thirst in an ocean of data” 90% 80% 20% of the world’s data of the world’s data amount of data was created in the today is traditional systems last two years unstructured leverage today © 2013 International Business Machines Corporation
    6. 6. Easy Question (LN (12,546,798*π )) ^ 2 / 34,576.46 0.00885 Select Payment where Owner = “David Jones” and Type (Product) = “Laptop” Owner Serial Number David Jones 45322190-AK Invoice # Vendor Payment INV10895 MyBuy $104.56Serial Number Type Invoice #45322190-AK LapTop INV10895 © 2013 International Business Machines Corporation
    7. 7. Hard QuestionComputer programs are natively explicit, fast and exacting in their calculation overnumbers and symbols….But Natural Language is implicit, highly contextual, ambiguousand often imprecise. Structured Where was X born? One day, from among his city views of Ulm, Otto chose a water color to send to Albert Einstein as a remembrance of Einstein´s birthplace. Unstructured X ran this? If leadership is an art then surely Jack Welch has proved himself a master painter during his tenure at GE. © 2013 International Business Machines Corporation
    8. 8. Informed Decision Making: Search vs. Expert Q&A Decision Maker Has Question Search Engine Distills to 2-3 Keywords Finds Documents containing Keywords Reads Documents, Finds Answers Expert Delivers Documents based on Popularity FindsDecision Evidence & Analyzes Maker Understands Question Asks NL Question Produces Possible Answers & Evidence Analyzes Evidence, Computes Confidence Considers Answer & Evidence Delivers Response, Evidence & Confidence8 © 2013 International Business Machines Corporation
    9. 9. Why is Jeopardy! so Difficult?answering complex natural language questions requires more than keyword evidence In May 1898 Portugal celebrated the In May 1898 Portugal celebrated the In May, Gary arrived in India In May, Gary arrived in India 400th anniversary of this explorer’s 400th anniversary of this explorer’s after he celebrated his after he celebrated his arrival in India arrival in India anniversary in Portugal anniversary in Portugal Legend Keyword “Hit” arrived in Reference Text celebrated celebrated Answer Red Text Weak evidence In May In May 1898 400th anniversary anniversary This evidence suggests Portugal in Portugal “Gary” is the answer BUT the system must learn that arrival in keyword matching may be weak relative to other India India types of evidence explorer Gary © 2013 International Business Machines Corporation
    10. 10. What It Takes to compete against Top Human Jeopardy! Players Each dot – actual historical human Jeopardy! games Top human players are remarkably good. Winning Human Winning Human Performance Performance Grand Champion Grand Champion Human Human Performance Performance 2007 QA Computer System 2007 QA Computer System More Confident More Confident Less Confident Less Confident © 2013 International Business Machines Corporation10
    11. 11. Levering Algorithms for Deeper EvidenceIn May 1898 Portugal celebrated the In May 1898 Portugal celebrated the On the 27th of May 1498, Vasco da On the 27th of May 1498, Vasco da400th anniversary of this explorer’s 400th anniversary of this explorer’s Gama landed in Kappad Beach Gama landed in Kappad Beacharrival in India. arrival in India. Legend Temporal Reasoning Statistical Paraphrasing celebrated landed in GeoSpatial Reasoning Portugal Reference Text Answer May 1898 400th anniversary Date 27th May 1498 Match Stronger evidence can be much Statistic harder to find and score… al Para- arrival in phrases  Search far and wide  Explore many hypotheses Geo-  Find judge evidence Spatia India l Kappad Beach  Many inference algorithms Reaso ning …and the evidence is still not Vasco da 100% certain explorer Gama © 2013 International Business Machines Corporation
    12. 12. Watson is a Massively Parallel Probabilistic Evidence-Based Architecture DeapQA generates and scores many hypotheses using an extensible collection if Natural Language Processing, Machine Learning and Reasoning Algoritms. These gather and weigh evidence over both structured and unstructured content to determine the answer with the best confidence Learned Models help combine and weigh the Evidence Evidence Sources Answer Models Models Sources Deep Inquiry Answer Evidence Models Models Evidence Scoring Retrieval Scoring Primary Candidate Models Models Search Answer GenerationInquiry/Topic Inquiry Hypothesis Hypothesis and Evidence Final Confidence Synthesis Analysis Decomposition Generation Scoring Merging & Ranking Hypothesis Hypothesis and Evidence Generation Scoring Responses with Confidence © 2013 International Business Machines Corporation
    13. 13. DeepQA: Incremental Progress in Answering Precisionon the Jeopardy Challenge: 6/2007-11/2010 IBM Watson Playing in the Winners Cloud v0.8 11/10 V0.7 04/10 v0.6 10/09 v0.5 05/09 v0.4 12/08 v0.3 08/08 v0.2 05/08 v0.1 12/07 Baseline 12/06 © 2013 International Business Machines Corporation
    14. 14. Healthcare & DataIBM Confidential © 2013 IBM
    15. 15. Our Watson Healthcare strategy solves 3 problems in clinical practiceRelated 3 problems medicine is a science but practiced as an art Impossible to keep up and have access to existing knowledge The number of untapped information that can be used as a source of knowledge is growing exponentially © 2013 International Business Machines Corporation
    16. 16. Our Watson Healthcare strategy solves 3 problems in clinical practice medicine is a science but practiced as an art  Estimated 30-40% of care in UK not based on available scientific evidence Grol, R. and Grimshaw, J. (2003)  5 year gap between publication of guidelines and changes in routine practice in Western healthcare systems, Lomas et al (1993)  1 out of 5 diagnoses are wrong  Unprecedented research commissioned by the EU has found that 23% of EU citizens have been a victim or the member of a family who has been a victim of a “serious medical error in a local hospital” or a “serious medical error from a medicine that was prescribed by a doctor”.  In all, only 17% of Austrians and Germans said that hospital patients were very likely or fairly likely to be able to avoid a serious medical error. © 2013 International Business Machines Corporation
    17. 17. Our Watson Healthcare strategy solves 3 problems in clinical practice Impossible to keep up and have access to existing knowledge  medical knowledge doubles every five years  81% of the physicians in the US report spending 5 hours or less a month reading medical journals  Medicine has become too complex and only 20% of the knowledge clinicians use is evidence based © 2013 International Business Machines Corporation
    18. 18. Our Watson Healthcare strategy solves 3 problems in clinical practice The number of untapped information that can be used as a source of knowledge is growing exponentially  16000 Hospitals worldwide collect data  80% of the data is unstructured and stored in hundred of forms such as lab results, images and medical transcripts  data will grow 800% over the next five years  90% of the digital data has been generated in the last 2 years  unstructured data will grow 50 times faster then structured data  patient monitoring equipment pumps out on average 1000 readings per second or 86400 reading a day  Data is getting more social. . .  20M articles on Wikipedia, 30B pieces of Facebook content are shared monthly  There are 156M public blogs, 12 terabites on tweets generates every day  70 percent of physicians report that at least one of their patients is sharing health measurement data with them © 2013 International Business Machines Corporation
    19. 19. Big Data: this is just the beginning 100 9000 Sensors & Devices 8000 Percentage of Percent of uncertain data 80 uncertain data Volume in Exabytes 7000 60 6000 Social Media 5000 40 You are here 4000 VoIP 20 3000 Enterprise Data 0 2010 2015Source: IBM Global Technology Outlook - 2012 © 2013 International Business Machines Corporation
    20. 20. Healthcare industry is beset with some of the most complexinformation challenges we collectively face “Medicine has become too complex. Only about 20% of the knowledge clinicians use today is evidence-based.” Steven Shapiro, Chief Medical & Scientific Officer, UPMC Steven Shapiro, Chief Medical & Scientific Officer, UPMC © 2013 International Business Machines Corporation
    21. 21. Watson in HealthcareOncology AdvisorIBM Confidential © 2013 IBM
    22. 22. NEJM Medical Concept Annotations – Attribute extractions Diseases Symptoms Medications Modifiers © 2013 International Business Machines Corporation
    23. 23. Putting the proper pieces together at the point of impactcan be life changing Fammpto Me . Histo y Pat Histor Sy dic Fin tions . a ry din difficulty swallowing fever ms Diagnosis Models Confidence gs dry mouth Symptoms Patient thirstSymptoms FamilyMedications Findings anorexia frequent urination dizziness Renal Failure History History no abdominal pain no back pain no cough UTI no diarrhea Diabetes Oral cancer History Family Bladder cancer Hemochromatosis Influenza Purpura Graves’ Disease (Thyroid Autoimmune) Hypokalemia cutaneous lupus Findings Medications History Patient osteoporosis hyperlipidemia Esophagitis frequent UTI hypothyroidism Alendronate pravastatin levothyroxine hydroxychloroquine urine dipstick: leukocyte esterase supine 120/80 mm HG heart rate: 88 bpm urine culture: E. Coli © 2013 International Business Machines Corporation
    24. 24. Watson in Healthcare Project Goals Build an intelligence engine to provide patient-specific diagnostic test and treatment recommendations Provide actionable treatment recommendations Built on the cognitive computing technologies developed in Watson by IBM Research Developed and Trained in collaboration with partners who are experts in their domain © 2013 International Business Machines Corporation
    25. 25. Working Together to Beat CancerCancer is an insidious disease and the second highest cause of death 1 in 4 3X individuals will die from cancer rate cancer cost climbs vs. std. health costs or 15-18% / yr. X 20% 263.8B of cancer cases receive the overall costs of cancer in wrong diagnosis initially with the US in 2010 some as high as 44% $$$$$$$$$$ ✔ ✔ ✔ ✔ ✔ ✔ ✔ $$$$$$$$$$ + + IBM $$$$$$$$$$ Working Together to Beat Cancer Source: American Cancer Society, National Health Institute © 2013 International Business Machines Corporation
    26. 26. Creating a Corpus of Knowledge for Cancer Care Ingestion of NCCN guidelines for breast cancer and lung cancer:  Roughly 500,000 unique combinations of breast cancer patient attributes.  Roughly 50,000 unique combinations of lung cancer patient attributes. Over 600,000 pieces of evidence ingested, from 42 different publications/publishers, including:  The Breast Journal, National Comprehensive Cancer Network (Clinical Practice Guidelines, Drug and Biologics compendium, et al.), American Journal Of Hematology, Annals Of Neurology, CA: A Cancer Journal For Clinicians, Cancer Journal, Cochrane, EBSCO, Hematological Oncology, Hepatology, International Journal Of Cancer, Journal Of Gene Medicine, Journal of Clinical Oncology, Journal of Oncology Practice, Massachusetts Medical Society Journal Watch, Massachusetts Medical Society New England Journal Of Medicine, Merck, Nephrology, UptoDate, Clinical Lung Cancer, Current Problems in Cancer, Cancer Treatment Reviews, Elseviers Monographs in Cancer (multiple), Clinical Breast Cancer, European Journal of Cancer, Lung Cancer (the journal). Watson has received 14,700 hours of training from clinicians Accurate: in the cases run, its 90% accurate, the goal is 100% accurancy, today physicians are about 50% accurate. IBM Confidential © 2013 International Business Machines Corporation
    27. 27. 27 © 2013 International Business Machines Corporation
    28. 28. 28 © 2013 International Business Machines Corporation
    29. 29. 29 © 2013 International Business Machines Corporation
    30. 30. 30 © 2013 International Business Machines Corporation
    31. 31. 31 © 2013 International Business Machines Corporation
    32. 32. Watson’s Five Core Capabilities Analyzes large volumes of unstructured Combines large amounts of unstructured data and structured data with structured data to be analyzed together Understands ambiguous and imprecise Interprets and understands natural questions using sophisticated natural language language questions algorithms Generates and evaluates hypotheses and Identifies many answers to questions with quantifies confidence in answers evidence to "explain" rationale for answers Supports iterative Enables iterative and interactive question and dialogue to refine results answering to refine and improve results Adapts and learns to Learns from additional evidence, additional questions and mistakes to improve accuracy improve results over time over time © 2013 International Business Machines Corporation
    33. 33. “ The application of what we know will have a bigger impact than any drug  or technology likely to be introduced in the next decade.” Sir Muir Gray,Director NHS National Knowledge Service & NHS Chief Knowledge Officer © 2013 International Business Machines Corporation
    34. 34. © 2013 International Business Machines Corporation

    ×