Commonsense knowledge for Machine Intelligence - part 3
Part 3: Applications and Open Issues
Commonsense for Machine Intelligence: Text to
Knowledge and Knowledge to Text
• A smart city is an urban area that uses different types of electronic
data collection sensors to supply information used to manage assets
and resources efficiently [Source: Wiki]
• Energy efficiency, ubiquitous information access, pollution control,
transparent government, good healthcare etc.
Smart City Characteristics
[Source: TU Wien 2015]
Potential Role of Commonsense Knowledge
• Smart Environment: Commonsense & related
concepts in pedestrian and traffic systems
• Smart Government: Commonsense in mining
urban legislation for transparency in govt.
• Smart Living: Opinion mining with commonsense
from social media data on health, safety etc.
• Smart Mobility: Autonomous vehicles with
• Smart People: Systems entailing commonsense
for 21st century education
• Smart Economy: Use of commonsense & world
knowledge on economics in policy decisions
[Note: These are just a few examples]
Commonsense for Smart Environment
• Buses in Barcelona run on routes
designed to optimize energy efficiency
• Street lights in Amsterdam brighten &
dim based on pedestrian usage
• Make such systems smarter using
commonsense knowledge & reasoning
• This enhances Smart Environment
[Source: TU Wien 2015]
Commonsense for Smart Government
• Ordinances: Laws passed &
enacted at local levels
• Mine Web-based ordinance data
to analyze issues addressed etc.
• Commonsense to select relevant
ordinances, interpret content,
map ordinance to dept. etc.
• Answer questions, e.g., Which
dept. was most active in a given
• This contributes to Smart Govt.
A Local Law in relation to the date of submission by the Mayor of the
proposed executive budget and budget message, the date of submission by
the Borough Presidents of recommendations in response to the Mayor
executive budget, the date of publication of a report by the director of the
independent budget office analyzing he executive budget, the date by
which the Council hearings pertaining to the executive budget shall
conclude, the date by which if the expense budget has not been adopted,
the expense budget and tax rate adopted as modified for the current fiscal
year shall be deemed to have......
[Source: NYC Council Websites 2017; Du et al., 2017]
Commonsense for Smart Living
• Opinion mining from social media, e.g., Twitter
• Commonsense knowledge for simulating human
judgement in mining tweets
• Address health & safety issues w.r.t. Smart Living
• E.g., Indonesian Peat Fires affected Singapore etc.
• Pollutants PM2.5 (particulate matter, diameter < 2.5µm)
harmful to lungs
• Public tweeted on pollutant impacts & controls
• Sentiment analysis for user satisfaction on health effects
• Commonsense useful in tweet selection & analysis
[Source: Forsyth T. 2014; Du, Emebo et al. 2016]
Commonsense for Smart Mobility
• If autonomous vehicles have
commonsense, they can make
intuitive, logical decisions
• This would avoid accidents e.g.
Tesla Model S & truck in 2016
• Tesla confused truck with
overpass due to height & tried
going through it, crashing!
• Commonsense can augment
domain KBs to make autonomous
• This promotes Smart Mobility
[Source: Driggs-Campbell et al. 2014, Persaud et al. 2017]
Commonsense for Smart People
• 21st century education affected by technology
• Computer literacy is crucial
• Commonsense in relevant systems is helpful
• Intelligent Tutoring Systems: more interactive;
cater to different expertise levels
• Writing aids in multiple languages: address
• This promotes inclusiveness & fosters global
education, enhancing Smart People
[Source: Long et al. 2011, Park et al. 2008, Koedinger et al., 2013]
Commonsense for Smart Economy
• Use commonsense with principles of economics
for depoliticized transparent policy
• Balance productivity & energy efficiency using
commonsense knowledge of systems
• Economic principle: “Free cooling can be used to
reduce air conditioning in data centers”
• Commonsense fact: “Provided location details are
taken into account”
• Economic principle: “More spending on public
schools will improve student performance”
• Commonsense fact: “As long as that spending is
done in the classroom”
[Source: Leef et al. 2007, Pawlish et al. 2014]
Further Research on Smart Cities & Commonsense
• Optimize data centers for
• Embed more commonsense in
street systems for smarter
• Build better healthcare apps for
• Enhance ITS, writing aids with
commonsense facts, collocations
for smarter people
• Make autonomous vehicles at
least as good as human drivers
for smarter mobility
• Improve overall urban legislation
policies for smarter government
• Driggs-Campbell K., Shia V., Bajcsy R., “Decisions for autonomous vehicles: Integrating sensors, communication,
and control”, ACM HiCoNS 2014, pp. 59-60.
• Du, X., Liporace, D. and Varde, A., “Urban Legislation Assessment by Data Analytics with Smart City
Characteristics”, IEEE UEMCON 2017, To appear.
• Du, X., Emebo, O., Varde, A., Tandon, N., Nag Chowdhury, S. and Weikum, G. Air quality assessment from social
media and structured data: Pollutants and health impacts in urban planning, IEEE ICDE Health Data Mgmt. &
Mining Workshop 2016, pp. 54-59.
• Forsyth, T. “Public concerns about transboundary haze: a comparison of Indonesia, Singapore and
Malaysia.”, Global Environmental Change 2014, 25:76-86.
• Koedinger, K., Brunskill, E., Baker, R., McLaulin, E. and Stamper, J., New Potentials for Data-Driven Intelligent
Tutoring System Development and Optimization, AI Magazine 2013, 34(3): 27-28.
• Leef, G., Smart Economics: Commonsense Answers to 50 Questions about Government, Taxes, Business and
Households, Foundation for Economic Education (FEE) 2007.
• Long, P., Siemens, G., Penetrating the fog: Analytics in learning & education, Educause Review 2011, 46(5): 31-40.
• New York City Council - Legislation, Legistar.council.nyc.gov, 2017
• Park, T., Lank, E., Poupart, P. and Terry, M., Is the sky pure today - awkchecker: An assistive tool for detecting and
correcting collocation errors. ACM Symposium on User Interface Software and Technology 2008, pp. 121–130.
• Pawlish, M., Varde, A., Robila, S. and Ranganathan, R., A call for energy efficiency in data centers, SIGMOD Record
2014, 43(1): 45-51.
• Persaud, P., Varde, A. and Robila, S., “Enhancing Autonomous Vehicles with Commonsense: Smart Mobility in
Smart Cities”, IEEE ICTAI Smart Cities Workshop 2017, To appear.
• Vienna University of Technology (TU Wien), European Smart Cities, Tech. Rep., Vienna, Austria 2015. 12
• Part 1: Acquiring Commonsense Knowledge
• What is commonsense
• Representing commonsense knowledge (CSK)
• Processes in acquisition of CSK
• Evaluation methods
• Part 2: Detecting and Correcting Odd Collocations in Text
• Meaning of collocations
• Treatment of odd collocations
• Linguistic classification approaches
• Tools for detection and correction
• Part 3: Applications and Open Issues
• Smart cities and their characteristics
• Role of commonsense to enhance smart city characteristics
• Specific applications with scope for further research 13
• Commonsense has made people smarter J
• Commonsense is making machines smarter!
• Commonsense will make smart cities smarter…