Filed in Programs Subscribe to Decision Science News by Email (one email per week, easy unsubscribe)
APPLICATION DEADLINE: APRIL 12, 2019
Machine learning is impacting the business world and the business research community. The CMU Summer Workshop on Machine Learning is intended to introduce junior researchers to the cutting-edge machine learning methods and their applications in the marketing and information systems research, and open up common ground for future discussions among researchers who are using or wish to use machine learning methods in their research.
The format of the workshop is a combination of lecture-style sessions, hands-on tutorials (directed primarily at doctoral students) and panel presentations. The lecture-style sessions and hands-on tutorials will cover the following topics:
Supervised Learning
Unsupervised Learning
Graphical Models
Deep Learning
Reinforcement Learning
Text, Image, Audio, Video mining
Bayesian Machine Learning
Causal Inference
Who should participate: The workshop is primarily intended for Ph.D. students in Consumer Behavior, Information Systems, Quantitative Marketing, or related business disciplines. We also welcome junior faculty members with an interest in machine learning methods to participate in the workshop.
Student application deadline is April 12, 2019. We will admit a total of 50 students; accepted applicants will be notified by April 20, 2019. Submit student application.
If there are more than two students who apply from the same school, we may ask the school to rank the students. Due to limited slots, we guarantee top 2 will be accepted, but depending on demand others may or may not be accepted to the workshop. Priority may be given to ISMS members.
Registration deadline is April 26, 2019: Ph.D. Students $50, Faculty $200.
All attendees must cover their own travel expenses and hotel costs. Meals during the workshop will be provided.
Filed in Programs Subscribe to Decision Science News by Email (one email per week, easy unsubscribe)
DEADLINE MARCH 31, 2019
In 2006, the European Association for Decision Making (www.eadm.eu) was pleased to announce the creation of the Jane Beattie Scientific Recognition Award to honor the memory of our late colleague, Jane Beattie. The award is to be made every two years and is intended for researchers who have recently completed the first stages of their careers – defined operationally by those who are 5 to 10 years post-PhD (see below).
The award is bestowed in recognition of “innovation in decision research”, as broadly understood within the Subjective Probability, Utility and Decision Making (SPUDM) tradition. In practical terms, this means that candidates should submit to the committee (a) a statement of 1,000 words or less that makes the case for their innovation; (b) one paper in which the innovation is presented for a scientific audience; and (c) a copy of their curriculum vitae. Candidates should also provide a statement as to when, and from where they received their PhDs.
The winner will receive a prize of 1,000 EUR, a certificate, and be asked to make a presentation at SPUDM 2019 in Amsterdam (www.spudm2019.com).
To be eligible for this award, candidates must have completed their PhDs no sooner than 5 years before the end of the most recent SPUDM meeting and no more than 10 years before the same date. Thus, to be eligible for the seventh award that will be presented at SPUDM in August 2019, candidates should have received their PhDs between August 2007 and August 2012.
The papers will be evaluated by a committee appointed by the Board of EADM consisting of Andreas Glöckner (University of Cologne, chair), Mehdi Moussaid (MPI Berlin, previous award winner), Arndt Bröder (University of Mannheim), Peter Ayton (University of London), and Cilia Witteman (Radboud University Nijmegen). To be considered for this award, papers and statements should be submitted before March 31, 2019 to: andreas.gloeckner@uni-koeln.de
In it, we noted that if you follow a normal curve to 5/8ths of the height of the bell, you’ll be at the ± 1 standard deviation marks.
So just measure up 5/8ths (we’ve drawn gridlines to make that easy), drop a line down to the x-axis and read off the standard deviation (making a small adjustment if it’s not 0 centered).
Click to enlarge
If you don’t like estimating 5/8ths, you can also try to identify the inflection point in the bell curve: the place where it goes from bending down to bending up. That occurs at exactly ± 1 standard deviations.
Applying those heuristics here, we’d estimate the standard deviation on this chart to be about 25 … and it is 25!
Another way to do this is to identify the range that encompasses about 99% of the data and divide by 6. Here the measurements run from -80 to 80 or about 160 units. This divided by 6 is 26.67.
Filed in Encyclopedia ,R Subscribe to Decision Science News by Email (one email per week, easy unsubscribe)
KNOW THY INFLECTION POINT
Click to Enlarge
We have drawn a lot of sorry-looking normal distributions in our life. It’s a shape that’s hard to get down without a lot of practice.
Here’s a few tips that can make it easier.
Start by marking out standard deviations on the x axis from -3 to +3
At x=0, draw a point to be the top of the bell curve
At ± 1 SDs draw points at about 5/8ths of the height
At ± 2 SDs draw points at about 1/8th of the height
At ± 3.25 SDs draw points on the X axis
Now, to get the bends right, we exploit the following cool fact: There is an inflection point at ±1 standard deviation. That is,
At 1 standard deviation, the curve stops bending down and starts bending up.
Draw through those points and voila, normal distribution!
You might be wondering if these tips are approximations or exact. The heights of the points are approximate but within 1-2% of the exact values. The inflection point at ± 1 SD is real. That’s right, the second derivative is 0 at exactly ± 1 standard deviation.
Filed in Conferences Subscribe to Decision Science News by Email (one email per week, easy unsubscribe)
SUBMISSION DEADLINE 18 APRIL 2019
The 2019 IAREP/SABE conference on Economic Psychology and Behavioural Economics in Dublin, Ireland. The conference will be held on 1st-4th September, 2019.
You are invited to submit your extended abstract (max 1000 words) or full paper before April 18, 2019.
Filed in Jobs Subscribe to Decision Science News by Email (one email per week, easy unsubscribe)
APPLICATION DEADLINE MARCH 31, 2019
The University of Pennsylvania, Social and Behavioral Science Initiative (SBSI) seeks applicants for a postdoctoral fellowship position for the 2019/2020 academic year. Funding is guaranteed for one year with the possibility of renewal for an additional year.
SBSI is a new interdisciplinary initiative comprised of scholars within the School of Arts and Sciences interested in the study of human social behavior and decision making.
The position is designed for individuals who have recently obtained a PhD degree in psychology or a related behavioral science discipline. The position is intended as a springboard for excellent researchers to help them build and establish their own research program. We are particularly interested in applicants who will pursue collaborative research with more than one SBSI scholar.
Applicants should specify in their research statement how their work connects with the interests of faculty in the SBSI. SBSI faculty are located primarily in the psychology department, but also in SAS departments that share an interest in human behavior and decision making, including communication, criminology, linguistics, philosophy, and political science. Topics of interest of faculty include judgement and decision-making, morality and cooperation, social cognition and evolutionary and cultural origins of behavior.
Benefits
Fellows receive a competitive salary and health insurance plus a modest research and travel budget. Fellows also benefit from access to the greater community of academics and leading research facilities equipped with cutting-edge instrumentation all on an urban campus in a vibrant city. Fellows are invited to join regular working group meetings within their field, plus career development workshops aimed at young researchers. Funding is guaranteed for one year with the possibility of renewal for an additional year.
Eligibility & Application
Applicants must have formally completed all requirements of the PhD. Candidates must submit a research statement that identifies potential collaborative opportunities with SBSI faculty at Penn, along with a CV, and contact information for two referees by March 31, 2019.
All eligible and complete applications will be evaluated by the Selection Committee and judged based on scientific excellence and fit.
Please direct applications and questions to: stonerl@sas.upenn.edu
Penn adheres to a policy that prohibits discrimination on the basis of race, color, sex, sexual orientation, gender identity, religion, creed, national or ethnic origin, citizenship status, age, disability, veteran status, or any other legally protected class. Background check required after a conditional job offer is made. Consideration of the background check will be tailored to the requirements of the job.
Filed in Conferences Subscribe to Decision Science News by Email (one email per week, easy unsubscribe)
DEADLINE EXTENDED TO FEB 10, 2019
Call for Papers
IC2S2 brings together researchers in computational science, complexity, and social science, and provides a platform for new work in the field of computational social science.
Contributed abstracts are presented orally in parallel thematic sessions or as posters at the three day conference, which takes place at the University of Amsterdam in the Netherlands from July 18 to 20.
Regular abstract submission
IC2S2 solicits abstracts from researchers in the social sciences with a clear component of computation, simulation or data analysis or data science. This includes for example sociology, psychology, communication science, anthropology, media studies, political science, public health, and economics. In addition, contributions from computer science, data science, and computational science with real-world applications in the social sciences or related fields, are welcome. We emphatically welcome abstracts that try to integrate both components. This is not limited to empirical studies; more general theoretical contributions are also welcome.
Topics of interest include, but are not limited to, the following:
Network analysis of social systems
Large-scale social experiments
Agent-based or other simulations of social phenomena
Text analysis and natural language processing (NLP) of social phenomena
Cultural patterns and dynamics
Computational science studies (sociology of science)
Social news curation and collaborative filtering
Social media studies
Theoretical discussions in computational social science
Causal inference and computational methods for social science
Ethics in computational social sciences
Reproducibility in computational social science
Large scale infrastructure in computational social science
Novel digital data sources
Computational analyses for addressing societal challenges
Methods and analyses of observational social data
Computational social science research in industry
Submission guidelines
Contributions to the conference should be submitted via EasyChair at:
Please follow the extended abstract template guidelines for Word (ic2s2-word-template.docx) and LaTeX (ic2s2-latex-template.zip) for formatting instructions. Note that abstracts should be submitted as a PDF file no larger than 20MB. Submissions that exceed the 2-page limit (including figures and references) will be automatically rejected.
The extended abstract should include a title and a list of 5 keywords, but no authors’ names or affiliations. The abstract should outline the main contribution, data and methods used, results, and the impact of the work. Authors are encouraged to include one figure in their submission (the figure counts towards the page limit).
Please do not include authors’ names and affiliations in the submitted document, as peer review will be double blind. Each extended abstract will be reviewed by multiple members of the Program Committee, composed of experts in computational social science.
When submitting on EasyChair you will be asked to provide information about the authors and their affiliations and to include a one-sentence summary of the extended abstract (20-50 words). The summary will be used for assigning reviewers. You can indicate a preference for an oral presentation or a poster presentation, but your preference may not be honored in the final decision.
Submissions will be non-archival, and the presented work can be already published, in preparation for publication elsewhere, or ongoing research. Submission implies willingness to present a talk or poster at the conference.
Who is Richard Price? According to our former Prof Stephen Stigler “Roughly half of Bayes’s famous essay was written by Richard Price, including the Appendix with all of the numerical examples.”
In reviewing the history of Bayes’s theorem and theology, one might wonder what Reverend Bayes had to say about this, and whether Bayes introduced his theorem as part of a similar argument for the existence of God. But the good reverend said nothing on the subject, and his theorem was introduced posthumously as part of his solution to predicting the probability of an event given specific conditions. In fact, while there is plenty of material on lotteries and hyperbolic logarithms, there is no mention of God in Bayes’s “Essay towards Solving a Problem in the Doctrine of Chances,” presented after his death to the Royal Society of London in 1763
That is, people have used Bayes Theorem to argue for the existence of God, buy Reverend Bayes was not one of them.
Filed in Encyclopedia ,Ideas ,R Subscribe to Decision Science News by Email (one email per week, easy unsubscribe)
WIND SPEEDS UP HEAT LOSS
Twenty years ago, a German colleague asked me what the deal was with wind chill. I guess they didn’t have it in Germany. I explained it was an attempt to communicate how it feels when there is a low temperature combined with wind. But my colleague wanted to know how they get the values in charts like these:
If you look closely at the first chart, it gives the formula for wind chill. It looks pretty hairy but it’s just a function of two things: wind speed and temperature. The windchill in Fahrenheit is just:
(Note that windchill is only defined for temperatures below 50°F and wind speeds above 3 mph.)
But where does this equation come from? This brochure has the answer. Researchers put sensors on people’s faces and had them walk in wind tunnels:
During the human trials, 6 male and 6 female volunteers were placed in a chilled wind tunnel. Thermal transducers were stuck to their faces to measure heat flow from the cheeks, forehead, nose and chin while walking 3 mph on a treadmill. Each volunteer took part in four trials of 90 minutes each and was exposed to varying wind speeds and temperatures. The NWS Wind Chill Temperature (WCT) index uses advances in science, technology, and computer modeling to provide an accurate, understandable, and useful formula for calculating the dangers from winter winds and freezing temperatures. The index does the following:
Calculates wind speed at an average height of 5 feet, the typical height of an adult human face, based on readings from the national standard height of 33 feet, typical height of an anemometer
Is based on a human face model
Incorporates heat transfer theory based on heat loss from the body to its surroundings, during cold and breezy/windy days
Lowers the calm wind threshold to 3 mph
Uses a consistent standard for skin tissue resistance
Assumes no impact from the sun, i.e., clear night sky.
A common misunderstanding is to assume that the wind can make something colder than the outside temperature. That can’t happen. Wind just speeds up the chilling process.
Ok, this is all fine, but we wanted a chart that would 1) make it easier to see the effect of the wind speed and temperature and 2) show the wind chill effect for typical February weather in New York and Chicago. We coded up the following:
click to enlarge
Note that in this chart, the Y axis isn’t the temperature, it’s the difference between the outside temperature and the windchill: the number of degrees you need to subtract.
So, when the wind is 10 miles per hour (red line), and the temperature is 25 degrees (x axis value of 25), the effect of wind chill is to lower the perceived temperature by 10 degrees (y axis value of -10), but at the same temperature when the wind is 25 miles per hour (blue line), wind lowers perceived temperature by about 16 degrees.
A useful takeaway is that with average winds and average temperatures, the effect of windchill is to lower the perceived temperature by about 7-12 degrees.
(BTW, if you like this stuff, you might enjoy our post on the heat index.)
Decision Science News is a website about Behavioral Economics, Data Science, Statistics, Marketing, Management, Psychology, Computer Science, Medicine, Policy, & Law
Subscribe to Decision Science News by Email (one email per week, easy unsubscribe)
