Maximizing Human Potential

“There is a problem in education. Assessments don’t align with real-world outcomes and most data about the learning process is underutilized.”

The Socos Solution

There exists today an enormous amount of data on how students are actually learning, and workers are performing. Simultaneously, while there is an increasing awareness of the importance of “21st Century Skills” there is little done to reinforce these traits. Socos’ solution is to take naturalistic data as the basis for our assessments. We close the educational loop by providing relevant feedback to educators on what they can do to improve life outcomes.

Read full story
Marshmallow

In the often-cited Marshmallow Study, where children are given the choice between one marshmallow now and two later, it has been clearly demonstrated that early-life self-restraint is predictive of life-long measures of success. Another aspect to this study was conducted in which children, prior to being given their first marshmallow, were promised crayons or similar enticement, by an adult who did not deliver on their promise. In each case of this reneging on a promise, children in the study ate their first marshmallow right way. Children were literally trained in a few minutes of the study to take what was available now because they could not rely on a future promise.

Miller, Chen SES Health Study

The dream of educational technolgy Adaptive Learning Everyone wants a system which can automatically tailor itself to the specific needs of each individual student. Our solution to this is the Cognitive Graph, a unique approach to modeling the connections between ideas uniquely for each learner. The graph can map any conceptual domain for both individuals and larger populations of student, allowing instructions to be dynamically targeted to a student's current misconceptions.

Jamaica Study

These findings show that a simple psychosocial stimulation intervention in early childhood for disadvantaged children can have a substantial effect on labor market outcomes and can compensate for developmental delays

Want to learn more about Socos?

Services

One of the most fundamental challenges in teaching is peering inside students' heads and figuring out what they're thinking. While education is a field rich with data, obtaining high-quality data and processing them meaningfully and efficiently remains difficult. Whether in formal classes, individualized tutoring, or casual web queries, learners continually generate questions, comments, proposals, discussions and a multitude of other assessable work.

These constitute valuable assessment data for informing instructors’ professional judgment, but systematically analyzing them across multiple students and time-points demands attention and resources beyond what most teachers can spare. The quantity of possible data to track defies ambition. The vast majority is lost to any broader perspective for instructors, educational leaders, and decision-makers. Lessons go untried, assessments unvalidated, population trends undetected and teaching opportunities missed. Rather than constantly designing and administering new tests, education needs tools which can actually make intelligent use of existing data.

Kindersight

Assessing the linguistic environment of kindergarteners.

Interest in improving early childhood learning across school and home settings is colliding with movements to increase standardized testing at ever younger ages. While testing proponents are rightfully concerned about measuring children’s learning, their design and use carry many problems. Tests are valid only for the population and purpose for which they were designed; eliminating cultural bias from tests is extremely difficult; and tests are often designed as sequestered experiences stripped from authentic contexts. Standardization narrows the range for what is considered acceptable progress regardless of developmental variation, and testing is intrusive, displacing instruction which might yield better learning. Testing and schoolwork burdens start to compete with at-home learning, rather than complementing it.

 

College Learners

College for America at Southern New Hampshire University

We are in the early stages of a project with College for America at Southern New Hampshire University, which Wired Magazine named the 16th most innovative organization in the world, and the number one most innovative educational organization. We are working with Southern New Hampshire University to implement new approaches to direct assessment in their competency-based learning paradigm. This continues to build on our interest in continuous passive assessment, an innovative method for real-time assessment of unstructured student work. With College for America we plan to validate our assessment algorithms based on their ability to predict concrete outcomes such as persistence and progression, as well as mapping them to established program goals and competencies.

 
Human Capital

In the often-cited Marshmallow Study, where children are given the choice between one marshmallow now and two later, it has been clearly demonstrated that early-life self-restraint is predictive of life-long measures of success. Another aspect to this study was conducted in which children, prior to being given their first marshmallow, were promised crayons or similar enticement, by an adult who did not deliver on their promise. In each case of this reneging on a promise, children in the study ate their first marshmallow right way. Children were literally trained in a few minutes of the study to take what was available now because they could not rely on a future promise.

Natualitic Technology Interventions

The three co-founders of Socos have combined their expertise in Cognition, Education and Machine Learning to develop our core Cognitive AnalyticsTM platform. It combines a set of proprietary algorithms for discovering underlying causes for student judgments with our unique Cognitive GraphTM Technology for capturing relationships connecting ideas and thinkers. The graph infers core conceptual factors underlying domains from normative sources as well as the student's own work. These factors power the analytics, providing metrics of the actual semantic content of students' work, not just word counts or time on task. Using the Cognitive Graph, instructors can quickly and easily perceive unique patterns in the understanding of individual students on specific assignments, across assignments, across domains and across time. The Graph easily extends to groups of students, allowing automated discovery of trends within and across populations.

Focus on 21st Century Skills

While education is a field rich with data, obtaining high-quality data and processing them meaningfully and efficiently remains difficult.

  • Human Capital

    In the often-cited Marshmallow Study, where children are given the choice between one marshmallow now and two later, it has been clearly demonstrated that early-life self-restraint is predictive of life-long measures of success. Another aspect to this study was conducted in which children, prior to being given their first marshmallow, were promised crayons or similar enticement, by an adult who did not deliver on their promise. In each case of this reneging on a promise, children in the study ate their first marshmallow right way. Children were literally trained in a few minutes of the study to take what was available now because they could not rely on a future promise.

  • Natualitic Technology Interventions

    Similar findings of low-grade interventions having substantial impact were found in addressing health problems of Children from families of low socioeconomic status. Eight years after intervention youth who participated had significantly less symptoms underlying low-grade health problems than controls. (Miller, Chen, and colleagues)

  • Focus on 21st Century Skills

    Another example of feedback creating lasting outcomes occurred during an intervention in Kingston, Jamaica. Severely underprivileged toddlers were identified in a government program. For three years, social workers went to the homes of those kids once a month and intervened not with the children but with their parents. They talked to parents about nutrition and social motivational energy skills. Twenty years later, those kids now adults, are actually studied by a group of economists; UC Berkeley and University of Chicago. What they find is that they earn 25% more in that atmosphere is that it’s not just the intervention. It’s that last time infatuated, they are generally, economically; indistinguishable for the last they thought. In other words, they were able to effectively erase the fact that they were from this severely autonomous population.

 

"Dr. Vivien Ming at SXSWEDU 2014"

Keeping the Promise of Educational

Research

  • circle
    Ming & Ming

    (2012)

    Modeling student conceptual knowledge from unstructured data using a hierarchical generative model. NIPS2012 Workshop: Personalizing Education With Machine Learning. South Lake Tahoe, CA.

    circle
    Ming & Ming

    (2012)

    Predicting Student Outcomes from Unstructured Data. IARIA, Data Analytics.

    circle
    Carlson, Ming & DeWeese

    (2012)

    Sparse codes for speech predict spectrotemporal receptive fields in the inferior colliculus. PLoS CompBio

    circle
    Ming & Ming

    (2012)

    Predicting Student Outcomes from Unstructured Data. UMAP2012.

  • circle
    Bumbacher & Ming

    (2012)

    Heirachical coding of natural signals in a dynamical system model. Cosyne2012

    circle
    Bumbacher & Ming

    (2012)

    Pitch-sensitive components emerge from hierarchical sparse coding natural sounds. ICPRAM2012

    circle
    Ming, N.C. & Baumer, E.P.S.

    (2011)

    Using text mining to characterize online discussion facilitation. Journal of Asynchronous Learning Networks, 15(2).

    circle
    Carlson, N., Ming, V.L. & DeWeese, M.R.

    (2010)

    A Sparse Representation of Speech Data. Sensory Coding & the Natural Environment, Gordon Research Conference.

  • circle
    Ming, V.L. & Holt

    (2009)

    Evidence of efficient coding in human speech perception. JASA 129, Num. 3: 1312-1321.

    circle
    Ming, N.C.

    (2009)

    Analogies vs. contrasts: A comparison of their learning benefits. In B. Kokinov, D. Gentner, & K. Holyoak (Eds.), New frontiers in analogy research: Proceedings of the second international conference on analogy (pp. 338-347). Sofia, Bulgaria: New Bulgarian University.

    circle
    Smith & Lewicki

    (2006)

    Efficient auditory coding. Nature 439, Num. 7079.

    circle
    Chang, N.M.

    (2006)

    Learning to Discriminate and Generalize through Problem Comparisons. Unpublished doctoral dissertation, Carnegie Mellon University, Pittsburgh PA.

  • circle
    Ming & Ming

    (2012)

    Modeling student conceptual knowledge from unstructured data using a hierarchical generative model. NIPS2012 Workshop: Personalizing Education With Machine Learning. South Lake Tahoe, CA.

    circle
    Ming & Ming

    (2012)

    Predicting Student Outcomes from Unstructured Data. IARIA, Data Analytics.

    circle
    Carlson, Ming & DeWeese

    (2012)

    Sparse codes for speech predict spectrotemporal receptive fields in the inferior colliculus. PLoS CompBio

    circle
    Ming & Ming

    (2012)

    Predicting Student Outcomes from Unstructured Data. UMAP2012.

  • circle
    Bumbacher & Ming

    (2012)

    Heirachical coding of natural signals in a dynamical system model. Cosyne2012

    circle
    Bumbacher & Ming

    (2012)

    Pitch-sensitive components emerge from hierarchical sparse coding natural sounds. ICPRAM2012

    circle
    Ming, N.C. & Baumer, E.P.S.

    (2011)

    Using text mining to characterize online discussion facilitation. Journal of Asynchronous Learning Networks, 15(2).

    circle
    Carlson, N., Ming, V.L. & DeWeese, M.R.

    (2010)

    A Sparse Representation of Speech Data. Sensory Coding & the Natural Environment, Gordon Research Conference.

  • circle
    Ming, V.L. & Holt

    (2009)

    Evidence of efficient coding in human speech perception. JASA 129, Num. 3: 1312-1321.

    circle
    Ming, N.C.

    (2009)

    Analogies vs. contrasts: A comparison of their learning benefits. In B. Kokinov, D. Gentner, & K. Holyoak (Eds.), New frontiers in analogy research: Proceedings of the second international conference on analogy (pp. 338-347). Sofia, Bulgaria: New Bulgarian University.

    circle
    Smith & Lewicki

    (2006)

    Efficient auditory coding. Nature 439, Num. 7079.

    circle
    Chang, N.M.

    (2006)

    Learning to Discriminate and Generalize through Problem Comparisons. Unpublished doctoral dissertation, Carnegie Mellon University, Pittsburgh PA.

  • circle
    Ming & Ming

    (2012)

    Modeling student conceptual knowledge from unstructured data using a hierarchical generative model. NIPS2012 Workshop: Personalizing Education With Machine Learning. South Lake Tahoe, CA.

    circle
    Ming & Ming

    (2012)

    Predicting Student Outcomes from Unstructured Data. IARIA, Data Analytics.

    circle
    Carlson, Ming & DeWeese

    (2012)

    Sparse codes for speech predict spectrotemporal receptive fields in the inferior colliculus. PLoS CompBio

    circle
    Ming & Ming

    (2012)

    Predicting Student Outcomes from Unstructured Data. UMAP2012.

  • circle
    Bumbacher & Ming

    (2012)

    Heirachical coding of natural signals in a dynamical system model. Cosyne2012

    circle
    Bumbacher & Ming

    (2012)

    Pitch-sensitive components emerge from hierarchical sparse coding natural sounds. ICPRAM2012

    circle
    Ming, N.C. & Baumer, E.P.S.

    (2011)

    Using text mining to characterize online discussion facilitation. Journal of Asynchronous Learning Networks, 15(2).

    circle
    Carlson, N., Ming, V.L. & DeWeese, M.R.

    (2010)

    A Sparse Representation of Speech Data. Sensory Coding & the Natural Environment, Gordon Research Conference.

  • circle
    Ming, V.L. & Holt

    (2009)

    Evidence of efficient coding in human speech perception. JASA 129, Num. 3: 1312-1321.

    circle
    Ming, N.C.

    (2009)

    Analogies vs. contrasts: A comparison of their learning benefits. In B. Kokinov, D. Gentner, & K. Holyoak (Eds.), New frontiers in analogy research: Proceedings of the second international conference on analogy (pp. 338-347). Sofia, Bulgaria: New Bulgarian University.

    circle
    Smith & Lewicki

    (2006)

    Efficient auditory coding. Nature 439, Num. 7079.

    circle
    Chang, N.M.

    (2006)

    Learning to Discriminate and Generalize through Problem Comparisons. Unpublished doctoral dissertation, Carnegie Mellon University, Pittsburgh PA.

Related Research


Areas of Interest

* Cognitive Development
* Computer-Mediated Learning
* Curriculum Development
* Educational Media
* Experimental Design In Education
* Learner-centered Education
* Mathematics Education
* Professional Development for Educators
* Science Education
* Simulation Learning Environments
* Teacher Education and Certification
* Technology and Schools

The three co-founders of Socos have combined their expertise

in Cognition, Education and Machine Learning.

About Us

Engin Bumbacher
Engin Bumbacher

Director of Research

Engin is devoted to the development of the company’s core cognitive modeling and predictive analytics technology. He did his master’s thesis project at the Redwood Center for Theoretical Neuroscience at UC Berkeley under the supervision of Dr. Vivienne Ming, applying and further developing elaborate models of information processing to human speech and music. Engin earned his master’s degree with honors in Neural Systems and Computation from the Swiss Federal Institute of Technology Zurich and the Institute of Neuroinformatics, both researching in the field of theoretical neuroscience and exploring models of collective intelligence through implementation of interactive flocking algorithms to control computer sound synthesis and 3D sound positioning. Prior to that, he finished his B.S. with honors in Physics at the same university.

Vivien Ming
Vivienne Ming

Executive Director

Dr. Vivienne Ming, named one of 10 Women to Watch in Tech in 2013 by Inc. Magazine, is a theoretical neuroscientist, technologist and entrepreneur. She is chief scientist at Gild, an inovative startup that applies machine learning to predict optimal candidates for technology jobs, and to bring meritocracy to job markets. She joined Gild in 2012 to oversee R&D and IP development, solving problems in data mining, text analysis, cognitive modeling and algorithm development. Dr. Ming also co-founded her own cutting-edge edtech startup, Socos, with her wife, Norma. She is a visiting scholar at UC Berkeley's Redwood Center for Theoretical Neuroscience pursuing her research in neuroprosthetics. In her free time, Dr. Ming also explores augmented cognition using technology like Google Glass and has been developing a predictive model of diabetes to better manage blood glucose levels. She sits on the boards of StartOut and Our Family Coalition and speaks on issues of LGBT inclusion and gender in technology. Previously, she was a junior fellow at Stanford’s Mind, Brain & Computation Center and earned her Ph.D. from Carnegie Mellon. Her work and research has received extensive media attention including the New York Times, NPR, Nature, O Magazine, Forbes, and The Atlantic.

Robert Doe
Norma Ming

Director of Learning Design

Dr. Norma Ming is a learning scientist and educational technology thought leader who works at the intersection of research and development, policy, and practice. She is a co-founder and the Director of Learning Design at Socos, which applies cognitive modeling to create adaptive, personalized educational technology. Dr. Ming merges a pragmatic understanding of the teaching enterprise with a long-term, systemic vision of how research can illuminate and policy can facilitate better learning. Her experience in teaching, professional development, assessment design, and curriculum evaluation crosses multiple disciplines and spans elementary through postgraduate students, teachers, administrators, and faculty trainers. Research projects have explored relationships among predictors, processes, and outcomes across a range of student populations and instructional models, from case studies to massive scale, individual or collaborative, with and without technology. Her policy advocacy highlights issues of equity in creating flexible paths and innovative resources to enable all learners to meet high expectations. Previously, she worked as Senior Research Scientist at the Nexus Research and Policy Center and taught as a lecturer in Education in Math, Science, and Technology at UC Berkeley’s Graduate School of Education, where she is now a visiting scholar. She earned an A.B. with honors in chemistry at Harvard University and a Ph.D. in cognitive psychology in the Program for Interdisciplinary Educational Research at Carnegie Mellon University.

CONTACT US

Learn more about Consulting services for educational datamining Developing a Cognitive Analytics plug-in for your course or LMS Designing adaptive learning systems

Recent Talks

* Lawrence Berkeley Labs
* LinkedIn
* SXSWEdu2014 keynote
* Pearson Foundation's "Be the Source" interviews
* EdLab Groundbreakers vialogue

Current Projects

* Bridging the Word Gap (US Prez.)
* KinderSight
* College Learners
* UT Austin
* Washington University