An Efficient Bit Vector Approach to Semantics-based Machine Perception in Resource-Constrained Devices
Download as PPTX, PDF1 like1,784 views
1) The document presents an efficient approach to implement machine perception on resource-constrained devices using bit vector encodings of semantic knowledge. 2) It describes how to translate sensor observations and prior semantic knowledge into bit vector representations to enable efficient explanation and discrimination of sensor readings. 3) The approach reduces the complexity of semantic reasoning from polynomial time to linear time, enabling machine perception capabilities on mobile devices.
An Efficient Bit Vector Approach to Semantics-based Machine Perception in Resource-Constrained Devices
- 1. 0101100011010011110010101100011011011010110001101001111001010110001101011000110100111 An Efficient Bit Vector Approach to Semantics-based Machine Perception in Resource-constrained Devices 0101100011010011110010101100011011011010110001101001111001010110001101011000110100111 Cory Henson, Krishnaprasad Thirunarayan, Amit Sheth Kno.e.sis – Ohio Center of Excellence in Knowledge-enabled Computing Wright State University, Dayton, OH, USA 1
- 2. The Patient of the Future MIT Technology Review, 2012 http://www.technologyreview.com/featuredstory/426968/the-patient-of-the-future/ 2
- 3. What if we could automate this sense making ability? … and do it efficiently and at scale 3
- 4. Sensing is a key enabler of the Internet of Things 50 Billion Things by 2020 (Cisco) BUT, how do we make sense of the resulting avalanche of sensor data? 4
- 5. People are good at making sense of sensory input What can we learn from cognitive models of perception? • The key ingredient is prior knowledge 5
- 6. Perception Cycle* Translating low-level signals Explanation into high-level knowledge 1 Observe Perceive Property Feature Prior Knowledge Focusing attention on those 2 aspects of the environment that Discrimination provide useful information * based on Neisser’s cognitive model of perception 6
- 7. To enable machine perception, Semantic Web technology is used to integrate sensor data with prior knowledge on the Web 7
- 8. The Web is becoming a global knowledge base 8
- 9. Prior knowledge on the Web W3C Semantic Sensor Network (SSN) Ontology Bi-partite Graph 9
- 10. Prior knowledge on the Web W3C Semantic Sensor Network (SSN) Ontology Bi-partite Graph 10
- 11. Explanation Explanation is the act of choosing the objects or events that best account for a set of observations; often referred to as hypothesis building Translating low-level signals Explanation into high-level knowledge 1 Observe Perceive Property Feature 11
- 12. Explanation Explanation is the act of choosing the objects or events that best account for a set of observations; often referred to as hypothesis building Inference to the best explanation • In general, explanation is an abductive problem; and hard to compute Finding the sweet spot between abduction and OWL • Single-feature assumption* enables use of OWL-DL deductive reasoner * An explanation must be a single feature which accounts for all observed properties 12
- 13. Explanation Explanatory Feature: a feature that explains the set of observed properties ExplanatoryFeature ≡ ∃ssn:isPropertyOf—.{p1} ⊓ … ⊓ ∃ssn:isPropertyOf—.{pn} Observed Property Explanatory Feature elevated blood pressure Hypertension clammy skin Hyperthyroidism palpitations Pulmonary Edema 13
- 14. Discrimination Discrimination is the act of finding those properties that, if observed, would help distinguish between multiple explanatory features Explanation Observe Perceive Property Feature Focusing attention on those 2 aspects of the environment that Discrimination provide useful information 14
- 15. Discrimination Expected Property: would be explained by every explanatory feature ExpectedProperty ≡ ∃ssn:isPropertyOf.{f1} ⊓ … ⊓ ∃ssn:isPropertyOf.{fn} Expected Property Explanatory Feature elevated blood pressure Hypertension clammy skin Hyperthyroidism palpitations Pulmonary Edema 15
- 16. Discrimination Not Applicable Property: would not be explained by any explanatory feature NotApplicableProperty ≡ ¬∃ssn:isPropertyOf.{f1} ⊓ … ⊓ ¬∃ssn:isPropertyOf.{fn} Not Applicable Property Explanatory Feature elevated blood pressure Hypertension clammy skin Hyperthyroidism palpitations Pulmonary Edema 16
- 17. Discrimination Discriminating Property: is neither expected nor not-applicable DiscriminatingProperty ≡ ¬ExpectedProperty ⊓ ¬NotApplicableProperty Discriminating Property Explanatory Feature elevated blood pressure Hypertension clammy skin Hyperthyroidism palpitations Pulmonary Edema 17
- 18. Our Motivation kHealth: knowledge-enabled healthcare Through physical monitoring and analysis, our cellphones could act as an early warning system to detect serious health conditions canary in a coal mine 18
- 19. How do we implement machine perception efficiently on a resource-constrained device? Use of OWL reasoner is resource intensive (especially on resource-constrained devices), in terms of both memory and time • Runs out of resources with prior knowledge >> 15 nodes • Asymptotic complexity: O(n3) 19
- 20. Approach 1: Send all sensor observations to the cloud for processing Approach 2: downscale semantic processing so that each device is capable of machine perception intelligence at the edge 20
- 21. Efficient execution of machine perception Use bit vector encodings and their operations to encode prior knowledge and execute semantic reasoning 010110001101 0011110010101 1000110110110 101100011010 0111100101011 000110101100 0110100111 21
- 22. Lifting and lowering knowledge Translate prior knowledge, observations, and explanations between SW and bit vector representation lower lift 22
- 23. Explanation: efficient algorithm INTUITION: The strategy employed relies on the use of the bit vector AND operation to discover and dismiss those features that cannot explain the set of observed properties. Observed Previous Current Property Prior Knowledge Explanatory Feature Explanatory Feature HN HM PE HN HM PE HN HM PE bp 1 bp 1 1 1 AND 1 1 1 => 1 1 1 cs 0 cs 0 1 0 pa 1 pa 1 1 0 AND => 1 1 0 23
- 24. Discrimination: efficient algorithm INTUITION: The strategy employed relies on the use of the bit vector AND operation to discover and assemble those features that discriminate between the explanatory features Observed Previous Current Discriminatin Property Prior Knowledge Explanatory Feature Explanatory Feature g Property HN HM PE bp 1 bp 0 1 1 HN HM PE HN HM PE bp 0 cs 0 cs 0 1 0 AND 1 1 0 => 0 1 0 cs 1 0 pa 1 pa 1 1 0 pa 0 … expected? = 0 1 0 => FALSE Is the property discriminating? ZERO Bit Vector … not-applicable? = => FALSE 0 0 0 24
- 25. Evaluation on a mobile device Efficiency Improvement • Problem size increased from 10’s to 1000’s of nodes • Time reduced from minutes to milliseconds • Complexity growth reduced from polynomial to linear O(n3) < x < O(n4) O(n) 25
- 26. 3 ideas to takeaway 1 Translate low-level data to high-level knowledge Machine perception can be used to convert low-level sensory signals into high-level knowledge useful for decision making 2 Prior knowledge is the key to perception Using SW technologies, machine perception can be formalized and integrated with prior knowledge on the Web 3 Intelligence at the edge By downscaling semantic inference, machine perception can execute efficiently on resource-constrained devices 26
- 27. Thank you. 0101100011010011110010101100011011011010110001101001111001010110001101011000110100111 An Efficient Bit Vector Approach to Semantics-based Machine Perception in Resource-constrained Devices 0101100011010011110010101100011011011010110001101001111001010110001101011000110100111 Cory Henson, Krishnaprasad Thirunarayan, Amit Sheth Kno.e.sis – Ohio Center of Excellence in Knowledge-enabled Computing 27
Editor's Notes
- #2: This is one example of an application being build from the research I will be discussingTitled: An efficient bit vector approach to semantics-based machine perception in resource-constrained devices
- #3: - Larry Smarr is a professor at the University of California, San DiegoAnd he was diagnosed with Chrones DiseaseWhat’s interesting about this case is that Larry diagnosed himselfHe is a pioneer in the area of Quantified-Self, which uses sensors to monitor physiological symptomsThrough this process he discovered inflamation, which led him to discovery of Chrones DiseaseThis type of self-tracking is becoming more and more common
- #4: - what if we could automate this sense making ability?- and what if we could do this at scale?
- #6: sense making based on human cognitive models
- #7: perception cycle contains two primary phasesexplanationtranslating low-level signals into high-level abstractions inference to the best explanationdiscriminationfocusing attention on those properties that will help distinguish between multiple possible explanationsused to intelligently task sensors and collect additional observations (rather than brute force approach of blindly collecting all observations)
- #12: perception cycle contains two primary phasesexplanationtranslating low-level signals into high-level abstractions inference to the best explanationdiscriminationfocusing attention on those properties that will help distinguish between multiple possible explanationsused to intelligently task sensors and collect additional observations (rather than brute force approach of blindly collecting all observations)
- #15: perception cycle contains two primary phasesexplanationtranslating low-level signals into high-level abstractions inference to the best explanationdiscriminationfocusing attention on those properties that will help distinguish between multiple possible explanationsused to intelligently task sensors and collect additional observations (rather than brute force approach of blindly collecting all observations)
- #19: - With this ability,many problems could be solved- For example: we could help solve health problems (before they become serious health problems) through monitoring symptoms and real-time sense making, acting as an early warning system to detect problematic health conditions
- #20: Intelligence distributed at the edge of the networkRequires resource-constrained devices (mobile phones, gateway notes, etc.) to be able to utilize SW technologies
- #21: Intelligence distributed at the edge of the networkRequires resource-constrained devices (mobile phones, gateway notes, etc.) to be able to utilize SW technologies
- #22: - compute machine perception inferences -- i.e., explanation and discrimination -- of high-complexity on a resource-constrained devices in miliseconds