• • • • •
Updates will be posted on this page, as well as emailed to the EE student mail list.
As always, use your best judgement and consider your own and others' well-being at all times.
Default Event Tag
In network data, relationships can be observed (e.g., as in a social network) or be built by
some similarity function (e.g., between data points in euclidean space). To find clusters
in such data, spectral clustering utilizes the eigenvectors of a similarity matrix, where the
(i, j)th element measures the similarity between points i and j. Unfortunately, the standard
spectral clustering algorithm fails when the similarity matrix is sparse (a common regime).
This talk will first discuss how a regularized spectral clustering algorithm can correct for
the problems created by this failure. The statistical improvements from regularization
are apparent in practice. The talk will theoretically characterize the improvement from
regularization under the degree corrected Stochastic Blockmodel. The talk will also discuss
contextualized spectral clustering in which the actors in the network have attributes that
correlate with the communities in the social network. We study the misclustering rate
of our proposed algorithm under a joint mixture model on the network and the node
covariates; this characterizes the algorithm as a statistical estimator. Applications with a
1,000,000 node DTI neuroconnectome and a 4,000,000 node online social network motivate
Cookies served at 3:45pm, 1st floor Lounge.
Good science, great technology and rapid innovation are the key attributes of a successful organization. Great organizations keep up with change by re-inventing themselves. Manufacturing frequently thought of as 'processes for making goods on a large scale' is erroneous and is dated.
Manufacture is rightfully value creation, and encompasses methods of realization post ideation. Volume flow within manufacturing is shifting towards mass customization. Batch size is shrinking, part identity, traceability, process flow, manufacturing geographies, supply chain all are taking on new meaning. Product foot-print is being redefined with the onset of wearables and IoT's. These are indeed wonderful times!
This presentation will begin with an insight to the start-up scene through the eyes of a venture partner. With a briefing on a few start-ups the author will provide insight on a few emerging technology pathways that could address needs in the manufacturing environment.
While the risk of traumatic brain injury (TBI) has recently become a health focus in athletics and militaries worldwide, data from fielding over 150,000 individually wearable sensors for monitoring hazardous events question common assumptions about the sources of those risks. This talk will review the development and fielding of a sensor system to record TBI events. Findings from two years of use by high-risk military and first responder groups will be discussed. A representative set of the over 5,000 recorded events will be reviewed, along with detailed event recreations. This will include data from the first recorded improvised explosive device (IED) attack and from training operations of civilian first responders. While the expected inertial and blast exposures are observed, the majority of the hazardous exposures are found to be unreported training events. These events will be discussed, along with future research directions.
Ultracold molecules at sub-microKelvin temperatures and trapped in crystals of light (optical lattices) present a new regime of physical chemistry and a new state of matter: complex dipolar matter. We present models for the quantum many-body statics and dynamics of present experiments on polar bi-alkali dimer molecules. We are developing and will discuss Hamiltonians and simulations for upcoming experiments on dimers beyond the alkali metals, including biologically and chemically important naturally occurring free radicals like the hydroxyl free radical (OH), as well as symmetric top polyatomic molecules like methyl fluoride (CH3F). These systems offer surprising opportunities in modeling and design of new materials, in addition to well-known exciting possibilities in quantum computing applications. For example, symmetric top polyatomics can be used to study quantum molecular magnets and quantum liquid crystals. Our numerical method of choice is massively parallel high performance computing via variational matrix-product-state (MPS) algorithms, a highly successful form of data compression used to treat lowly entangled dynamics and statics of many-body systems with large Hilbert spaces; we supplement our calculations with exact diagonalization and simpler variational, perturbative, and other approaches. We use MPS algorithms not only to produce experimentally measurable quantum phase diagrams but also to explore the dynamical interplay between internal and external degrees of freedom inherent in complex dipolar matter. Our group maintains open source code (openTEBD and openMPS) available freely and used widely. We have worked and will continue to work closely with experimentalists throughout our projects, and make detailed use of ultracold molecular properties and constants to provide concrete and accurate explanations, guidance, and inspiration.
 Kenji Maeda, M. L. Wall, and L. D. Carr, ''Hyperfine structure of OH molecule in electric and magnetic fields,'' New J. Phys., under review, arXiv:1410.3849 (2014)
 M. L. Wall, Kenji Maeda, and L. D. Carr, ''Realizing unconventional quantum magnetism with symmetric top molecules,'' New J. Phys., under review, arXiv:1410.4226 (2014)
 M. L. Wall, Kenji Maeda, and L. D. Carr, ''Simulating quantum magnets with symmetric top molecules,'' Ann. Phys. (Berlin) 525, 845 (2013)
 M. L. Wall, E. Bekaroglu and L. D. Carr, ''The Molecular Hubbard Hamiltonian: Field Regimes and Molecular Species,'' Phys. Rev. A, 88, 023605 (2013)
 M. L. Wall and L. D. Carr, ''Out of equilibrium dynamics with Matrix Product States,'' New J. Phys. 14, 125015 (2012)
 L. D. Carr, David DeMille, Roman V. Krems, and Jun Ye, ''Cold and Ultracold Molecules: Science, Technology, and Applications,'' New J. Phys. 11, 055049 (2009)
Tentative Abstract: The last 10-15 years have seen huge advances in the use of ultra-cold atoms techniques to study strongly correlated physics. The combination of laser cooling and optical lattices has allowed experimentalists to create clean, cold, and tunable artificial crystals of atoms and light. These systems provide an excellent platform in which to study fundamental underpinnings of materials physics involving degenerate Bose and Fermi gasses in rigid periodic potentials. I will discuss current experimental work in Lev Lab studying a multimode cavity QED system that gives rise to a novel kind of dynamical optical lattice which will generate a smecticly ordered superfluid state of 87Rb. Finally, I will discuss future work to generate glassy ordered states.