cognitive computing ein update rund um watson
TRANSCRIPT
Cognitive Computing –
Ein Update rund um Watson
Nicole Roik
Solution Sales Professional for Cognitive and Analytics
in Insurance
Executive IT Architect
Beobachten
Interpretieren
Evaluieren
Entscheiden
Lernen
Motorische Funktion
Planung
Vorstellung
Emotionalität
4
Welches sind die wesentlichen Bausteine von
Watson?
© 2017 International Business Machines Corporation6
Kognitive Systeme erschaffen eine neue Partner-
schaft zwischen Menschen und Technologie.
wo Menschen
brillieren:
Dilemmata
Mitgefühl
Träumen
Abstraktion
Vorstellungskraft
Moral
Generalisierung
wo kognitive
Systeme brillieren:
gesunder Menschenverstand
Wissen lokalisieren
Mustererkennung
Machine Learning
Vorurteile eliminieren
Beliebige Kapazität
Kreativität
© 2017 International Business Machines Corporation7
Beispiel für evidenzbasierte Diagnosefindung
•Extract Symptoms from record
•Use paraphrasings mined from text to handle alternate
phrasings and variants
Sym
pto
ms
difficulty swallowing
dizziness
anorexia
fever
dry mouth
thirst
frequent urination
Symptoms
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
© 2017 International Business Machines Corporation8
Beispiel für evidenzbasierte Diagnosefindung
•Extract Symptoms from record
•Use paraphrasings mined from text to handle alternate
phrasings and variants
•Perform broad search for possible diagnoses
•Score Confidence in each diagnosis based on evidence
so far
Most Confident Diagnosis: Influenza
Sym
pto
ms
UTI
Diabetes
Influenza
Hypokalemia
Renal failure
Diagnosis Models Confidence
difficulty swallowing
dizziness
anorexia
fever
dry mouth
thirst
frequent urination
Symptoms
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
© 2017 International Business Machines Corporation8
Beispiel für evidenzbasierte Diagnosefindung
•Extract Symptoms from record
•Use paraphrasings mined from text to handle alternate
phrasings and variants
•Perform broad search for possible diagnoses
•Score Confidence in each diagnosis based on evidence
so far
• Identify negative Symptoms
Most Confident Diagnosis: Diabetes
Sym
pto
ms
UTI
Diabetes
Influenza
Hypokalemia
Renal failure
no abdominal pain
no back pain
no cough
no diarrhea
Diagnosis Models Confidence
difficulty swallowing
dizziness
anorexia
fever
dry mouth
thirst
frequent urination
Symptoms
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
© 2017 International Business Machines Corporation8
Beispiel für evidenzbasierte Diagnosefindung
•Extract Symptoms from record
•Use paraphrasings mined from text to handle alternate
phrasings and variants
•Perform broad search for possible diagnoses
•Score Confidence in each diagnosis based on evidence
so far
• Identify negative Symptoms
•Reason with mined relations to explain away symptoms
(thirst is consistent w/ UTI)
Most Confident Diagnosis: UTI
Sym
pto
ms
UTI
Diabetes
Influenza
Hypokalemia
Renal failure
no abdominal pain
no back pain
no cough
no diarrhea
Diagnosis Models Confidence
difficulty swallowing
dizziness
anorexia
fever
dry mouth
thirst
frequent urination
Symptoms
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
© 2017 International Business Machines Corporation8
Beispiel für evidenzbasierte Diagnosefindung
•Extract Symptoms from record
•Use paraphrasings mined from text to handle alternate
phrasings and variants
•Perform broad search for possible diagnoses
•Score Confidence in each diagnosis based on evidence
so far
• Identify negative Symptoms
•Reason with mined relations to explain away symptoms
(thirst is consistent w/ UTI)
• Extract Family History
• Use Medical Taxonomies to generalize medical
conditions to the granularity used by the models
Most Confident Diagnosis: Diabetes
Sym
pto
ms
UTI
Diabetes
Influenza
Hypokalemia
Renal failure
no abdominal pain
no back pain
no cough
no diarrhea
(Thyroid Autoimmune)
Diagnosis Models Confidence
difficulty swallowing
dizziness
anorexia
fever
dry mouth
thirst
frequent urination
Fam
ily
His
tory
Graves’ Disease
Oral cancer
Bladder cancer
Hemochromatosis
Purpura
Family History
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
© 2017 International Business Machines Corporation8
Beispiel für evidenzbasierte Diagnosefindung
•Extract Symptoms from record
•Use paraphrasings mined from text to handle alternate
phrasings and variants
•Perform broad search for possible diagnoses
•Score Confidence in each diagnosis based on evidence
so far
• Identify negative Symptoms
•Reason with mined relations to explain away symptoms
(thirst is consistent w/ UTI)
• Extract Family History
• Use Medical Taxonomies to generalize medical
conditions to the granularity used by the models
Most Confident Diagnosis: UTI
Sym
pto
ms
UTI
Diabetes
Influenza
Hypokalemia
Renal failure
no abdominal pain
no back pain
no cough
no diarrhea
(Thyroid Autoimmune)
Diagnosis Models
frequent UTI
cutaneous lupus
hyperlipidemia
osteoporosis
hypothyroidism
Confidence
difficulty swallowing
dizziness
anorexia
fever
dry mouth
thirst
frequent urination
Fam
ily
His
tory
Graves’ Disease
Oral cancer
Bladder cancer
Hemochromatosis
Purpura
Pati
en
tH
isto
ry
Patient History
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
© 2017 International Business Machines Corporation8
Beispiel für evidenzbasierte Diagnosefindung
•Extract Symptoms from record
•Use paraphrasings mined from text to handle alternate
phrasings and variants
•Perform broad search for possible diagnoses
•Score Confidence in each diagnosis based on evidence
so far
• Identify negative Symptoms
•Reason with mined relations to explain away symptoms
(thirst is consistent w/ UTI)
• Extract Family History
• Use Medical Taxonomies to generalize medical
conditions to the granularity used by the models
•Extract Medications
•Use database of drug side-effects
•Together, multiple diagnoses may best explain symptoms
•Extract Findings: Confirms that UTI was present
Most Confident Diagnosis: Esophagitis
Sym
pto
ms
UTI
Diabetes
Influenza
Hypokalemia
Renal failure
no abdominal pain
no back pain
no cough
no diarrhea
(Thyroid Autoimmune)
Esophagitis
pravastatin
Alendronate
levothyroxine
hydroxychloroquine
Diagnosis Models
frequent UTI
cutaneous lupus
hyperlipidemia
osteoporosis
hypothyroidism
Confidence
difficulty swallowing
dizziness
anorexia
fever
dry mouth
thirst
frequent urination
Fam
ily
His
tory
Graves’ Disease
Oral cancer
Bladder cancer
Hemochromatosis
Purpura
Pati
en
tH
isto
ryM
ed
icati
on
s
Her medications were levothyroxine,
hydroxychloroquine, pravastatin, and
alendronate.
Medications
© 2017 International Business Machines Corporation8
Beispiel für evidenzbasierte Diagnosefindung
•Extract Symptoms from record
•Use paraphrasings mined from text to handle alternate
phrasings and variants
•Perform broad search for possible diagnoses
•Score Confidence in each diagnosis based on evidence
so far
• Identify negative Symptoms
•Reason with mined relations to explain away symptoms
(thirst is consistent w/ UTI)
• Extract Family History
• Use Medical Taxonomies to generalize medical
conditions to the granularity used by the models
•Extract Medications
•Use database of drug side-effects
•Together, multiple diagnoses may best explain symptoms
•Extract Findings: Confirms that UTI was present
Most Confident Diagnosis: UTI
Sym
pto
ms
UTI
Diabetes
Influenza
Hypokalemia
Renal failure
no abdominal pain
no back pain
no cough
no diarrhea
(Thyroid Autoimmune)
Esophagitis
pravastatin
Alendronate
levothyroxine
hydroxychloroquine
Diagnosis Models
frequent UTI
cutaneous lupus
hyperlipidemia
osteoporosis
hypothyroidism
Confidence
difficulty swallowing
dizziness
anorexia
fever
dry mouth
thirst
frequent urination
Fam
ily
His
tory
Graves’ Disease
Oral cancer
Bladder cancer
Hemochromatosis
Purpura
Pati
en
tH
isto
ryM
ed
icati
on
sF
ind
ing
s
supine 120/80 mm HG
urine dipstick:
leukocyte esterase
urine culture: E. Coli
heart rate: 88 bpm
Findings
A urine dipstick was positive for leukocyte esterase and
nitrites. The patient given a prescription fo ciprofloxacin
for a urinary tract infection. 3 days later, patient reported
weakness and dizziness. Her supine blood pressure was
120/80 mm Hg, and pulse was 88.
© 2017 International Business Machines Corporation8
Für Versicherungen entstehen spezifische
„Insurance Patterns”
Virtuelle Agenten im Kundenkontakt
leisten Dienste, führen und beraten
über Web oder Mobile
Transformation der
Kundeninteraktion
(Engagement)Anwendungen für Mitarbeiter, die
unterschiedliche Benutzer im
Unternehmen für Kundendienste
befähigen
Befähigte
Berater
Anwendungen für
Mitarabeiter, die Middle- und
Back-Office-Prozesse
optimieren
Optimierter
Betrieb
© 2017 International Business Machines Corporation16
© 2017 IBM Corporation
Watson liest und versteht Dokumente, sodass Einblicke in Relation geschaffen werden
19
Etappe 1: Show Case• Beispieldaten laden
• Watson Explorer Kernkompetenzen aufzeigen
Etappe 2: Erster
Prototyp, PoC• Kunden den Mehrwert von Watson
Explorer zeigen
• Zeiteffektiv
• Kostengünstig
• Prototyp nicht serienreif
Etappe 3: Produktionslevel•Produktionsbereit, lieferbar
•Löst echte Kundenprobleme
•Auf die Bedürfnisse des Kunden
abgestimmt
Etappe 4: Komplexe Vorgänge• Komplexe Kundenbedürfnisse werden erfüllt
• Produktionsbereit, lieferbar
• Löst echte Kundenprobleme
• Mehrere Systemintegrationen mit verschiedenen Systemen und Technologien
Nächste Schritte
© 2017 International Business Machines Corporation30
© 2017 IBM Corporation
Watson Developer Cloud - a set of cognitive capabilities
Rapidly prototype and build powerful cognitive apps in the cloud
Concept Expansion
Machine Translation
Message Resonance
Question and Answer
Personality Insights Relationship Extraction
Speech to Text
Text to speech Tradeoff AnalyticsVisualization Rendering Visual Recognition
Concept Insights
Analyzes the visual appearance of
images or video frames to
understand what is happening.
Translate text from one language to
another.
Personality insights to engage users
on their own terms
Graphical representations of data
analysis for easier understanding
Direct responses to users inquiries
fueled by primary document sources
Provides highly accurate, low latency
speech recognition capabilities
Synthesizes natural-sounding speech
from text
Communicate with people with a
style and words that suits them
Locate relevant documents that may
not directly mention your query.
Intelligently finds relationships
between sentences components
(nouns, verbs, subjects, objects, etc.)
Helps make better choices under
multiple conflicting goals with smart
visualizations and analytical
recommendations.
Maps euphemisms or colloquial
terms to more commonly understood
phrases
32
© 2017 IBM Corporation
Macht Watson dasselbe wie eine Suchmaschine?
IBM Watson
Bekommt Fragen gestellt
Search engine vs.
Leitet 2-3 Schlagworte ab
Liest Dokumente
Findet Antworten
Findet und analysiert Evidenz
Findet Dokumente nach Schlagworten
Findet besonders populäre Dokumente
Stellt Fragen
Versteht Fragen
Errechnet sichere Antworten
Gibt Antworten, Evidenz und sichere Belege
Beachtet Antworten und Antwortevidenz
Gibt mögliche Antworten und Evidenz dafür
Analysiert Evidenzen
How Watson Works Video https://www.youtube.com/watch?v=_Xcmh1LQB9I© 2017 International Business Machines Corporation
33