Mathematics of AI

Mathematics of AI Publications

2020

Decentralized stochastic non-convex optimization over weakly connected time-varying digraphs
Songtao Lu, Chai Wah Wu
IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP), 2020

Multimedia conferencing system for determining participant engagement
Dettori, Paolo and Schaffa, Frank Andre and Smilkov, Daniel and Westerink, Peter Hans and Wu, Chai Wah
US Patent 10,586,131
Abstract

2019

Free-form production based on causal predictive models
Lior Horesh, Chai W Wu, Ramesh Natarajan, Raya Horesh, David Nahamoo, Christopher Wildsmith, Michael Widman
US Patent App. 15/876,718

A Family of Robust Stochastic Operators for Reinforcement Learning
Yingdong Lu, Mark Squillante, Chai Wah Wu
NeurIPS 2019 : Thirty-third Conference on Neural Information Processing Systems, pp. 15626-15636
Abstract reinforcement learning, operator, mathematical optimization, computer science

Unification of Recurrent Neural Network Architectures and Quantum Inspired Stable Design
Murphy Yuezhen Niu, Lior Horesh, Michael O'Keeffe, Isaac Chuang
https://openreview.net/pdf?id=SJfFTjA5KQ, 2019

Quantum Control in the Eye of Artificial Neural Networks
Murphy Yuezhen Niu, Lior Horesh, Isaac Chuang
22nd Annual Conference on Quantum Information Processing, 2019

2018

Designing communication systems via iterative improvement: error correction coding with Bayes decoder and codebook optimized for source symbol error
Chai Wah Wu
arXiv preprint arXiv:1805.07429, 2018
Abstract pattern recognition, mathematics, linear code, hamming code, estimator, error detection and correction, decoding methods, codebook, bit error rate, bayes theorem, artificial intelligence

Can machine learning identify interesting mathematics? An exploration using empirically observed laws
Chai Wah Wu
arXiv preprint arXiv:1805.07431, 2018
Abstract nice, multiplicative function, mathematics, mathematical structure, machine learning, law, integer sequence, benford s law, artificial intelligence

ProdSumNet: reducing model parameters in deep neural networks via product-of-sums matrix decompositions
Chai Wah Wu
arXiv preprint arXiv:1809.02209, 2018
Abstract word error rate, small number, operator, matrix, mathematics, machine learning, mnist database, deep learning, artificial neural network, artificial intelligence, architecture

Composable and Versatile Privacy via Truncated CDP
Mark Bun, Cynthia Dwork, Guy N. Rothblum, Thomas Steinke
50th Annual ACM Symposium on the Theory of Computing (STOC), 2018

Bounded-Velocity Stochastic Control for Dynamic Resource Allocation
Gao, Xuefeng and Lu, Yingdong and Sharma, Mayank and Squillante, Mark S and Bosman, Joost W
arXiv preprint arXiv:1801.01221, 2018
Abstract

Stochastic Optimization Models for Workforce Planning, Operations, and Risk Management
Davis, Michael J and Lu, Yingdong and Sharma, Mayank and Squillante, Mark S and Zhang, Bo
Service Science 10(1), 40--57, INFORMS, 2018
Abstract

Heuristic policies for stochastic knapsack problem with time-varying random demand
Lu, Yingdong
Applied Stochastic Models in Business and Industry, Wiley Online Library, 2018
Abstract

Optimal Dynamic Control for Input-Queued Switches in Heavy Traffic
Yingdong Lu, Siva Theja Maguluri, Mark S. Squillante, Tonghoon Suk
2018 Annual American Control Conference (ACC), pp. 3804-3809
Abstract steady state, scheduling, schedule, scaling, real time computing, optimal control, linear programming, job shop scheduling, control engineering, computer science

An Optimal Scheduling Policy for the 2 X 2 Input-Queued Switch with Symmetric Arrival Rates
Y. Lu, S.T. Maguluri, M.S. Squillante, T. Suk, X. Wu
measurement and modeling of computer systems, pp. 217-223, 2018
Abstract scheduling, real time computing, queue, job shop scheduling, horizon, computer science

Heuristic policies for stochastic knapsack problem with timevarying random demand
Yingdong Lu
Applied Stochastic Models in Business and Industry 34(6), 1007-1016, 2018
Abstract revenue management, monotonic function, mathematical optimization, knapsack problem, heuristic, economics, econometrics, dynamic programming, bellman equation

Shared neural coding for social hierarchy and reward value in primate amygdala
J. Munuera, M. Rigotti, C.D. Salzman
Nature Neuroscience, 1, Nature Publishing Group, 2018
Abstract

Signatures and mechanisms of low-dimensional neural predictive manifolds
Stefano Recanatesi, Matthew Farrell, Guillaume Lajoie, Sophie Deneve, Mattia Rigotti, Eric Shea-Brown
bioRxiv, 2018
Abstract spatial memory, recurrent neural network, predictive learning, network dynamics, machine learning, experimental data, entorhinal cortex, curse of dimensionality, biology, bioinformatics, artificial neural network, artificial intelligence

The geometry of abstraction in hippocampus and prefrontal cortex
Silvia Bernardi, Marcus K Benna, Mattia Rigotti, Jerome Munuera, Stefano Fusi, Daniel Salzman
bioRxiv, 2018
Abstract prefrontal cortex, inference, hippocampus, geometry, deductive reasoning, cognition, biology, artificial neural network, anterior cingulate cortex, abstraction

Efficient ConvNets for Analog Arrays
Malte J Rasch, Tayfun Gokmen, Mattia Rigotti, Wilfried Haensch
arXiv:1807.01356, 2018
Abstract

Epidemic-Like Stochastic Processes with Time-Varying Behavior: Structural Properties and Asymptotic Limits
Lu, Yingdong and Squillante, Mark S and Wah Wu, Chai
ACM SIGMETRICS Performance Evaluation Review 45(2), 160--166, ACM, 2018
Abstract

Understanding Unequal Gender Classification Accuracy from Face Images
Vidya Muthukumar, Tejaswini Pedapati, Nalini Ratha, Prasanna Sattigeri, Chai-Wah Wu, Brian Kingsbury, Abhishek Kumar, Samuel Thomas, Aleksandra Mojsilovic, Kush R. Varshney
arXiv preprint arXiv:1812.00099, 2018
Abstract suspect, stereotype, pattern recognition, luminance, computer science, cognitive psychology, cheek structure, artificial neural network, artificial intelligence

A General Family of Robust Stochastic Operators for Reinforcement Learning
Lu, Yingdong and Squillante, Mark S and Wu, Chai Wah
arXiv preprint arXiv:1805.08122, 2018
Abstract

Record values in appending and prepending bitstrings to runs of binary digits
Chai Wah Wu
arXiv preprint arXiv:1810.02293, 2018
Abstract strings, mathematics, integer, discrete mathematics, binary number, algebra

Numerical Continuation of Invariant Solutions of the Complex Ginzburg-Landau Equation
V. Lopez
Communications in Nonlinear Science and Numerical Simulation 61, 2018

VISUALLY CONFIGURABLE PRIVACY ENFORCEMENT
Horesh Lior, Horesh Raya, Pistoia Marco, Tripp Omer
2018
enforcement, computer security, business

SYSTEM, METHOD, AND APPARATUS FOR CROWD-SOURCED GATHERING OF APPLICATION EXECUTION EVENTS FOR AUTOMATIC APPLICATION TESTING AND REPLAY
Horesh Lior, Horesh Raya, Pistoia Marco, Tripp Omer
2018
computer science

Sensory holograms
Aleksandr Y Aravkin, Lior Horesh, Raya Horesh, Dimitri Kanevsky, James J Wynne
US Patent 9,958,829

Experimental design for nonparametric correction of misspecified dynamical models
Gal Shulkind, Lior Horesh, Haim Avron
SIAM/ASA Journal on Uncertainty Quantification 6(2), 880--906, Society for Industrial and Applied Mathematics, 2018

System, method, and apparatus for crowd-sourced gathering of application execution events for automatic application testing and replay
Lior Horesh, Raya Horesh, Marco Pistoia, Omer Tripp
US Patent App. 15/222,067

Visually configurable privacy enforcement
Lior Horesh, Raya Horesh, Marco Pistoia, Omer Tripp
US Patent App. 15/243,278

Low Cost Container Condition Monitor
Amos Cahan, Guy M Cohen, Lior Horesh, Raya Horesh
US Patent App. 15/272,146

Predicting automobile future value and operational costs from automobile and driver information for service and ownership decision optimization
Amos Cahan, Guy M Cohen, Lior Horesh, Raya Horesh
US Patent App. 15/271,601

Multi-level modeling of steam assisted gravity drainage wells
Andrew R Conn, Lior Horesh, Matthias Kormaksson, Moshood Omolade Saliu, Theodore G Van Kessel
US Patent App. 15/275,951

Controlling operation of a steam-assisted gravity drainage oil well system by adjusting controls based on forecast emulsion production
Andrew R Conn, Lior Horesh, Matthias Kormaksson, Moshood O Saliu, Theodore G Van Kessel
US Patent App. 15/276,189

Controlling operation of a steam-assisted gravity drainage oil well system by adjusting controls to reduce model uncertainty
Andrew R Conn, Lior Horesh, Matthias Kormaksson, Moshood O Saliu, Theodore G Van Kessel
US Patent App. 15/276,168

Controlling operation of a steam-assisted gravity drainage oil well system utilizing continuous and discrete control parameters
Andrew R Conn, Lior Horesh, Mathias Kormaksson, Moshood O Saliu, Theodore G Van Kessel
US Patent App. 15/276,152

Cross-well allocation optimization in steam assisted gravity drainage wells
Andrew R Conn, Lior Horesh, Matthias Kormaksson, Moshood Omolade Saliu, Theodore G Van Kessel
US Patent App. 15/276,020

Controlling operation of a steam-assisted gravity drainage oil well system by adjusting multiple time step controls
Andrew R Conn, Lior Horesh, Matthias Kormaksson, Moshood O Saliu, Theodore G Van Kessel
US Patent App. 15/276,178

Unmanned vehicle for servicing equipment
Guy M Cohen, Lior Horesh, Raya Horesh, Jonathan Lenchner
US Patent App. 15/406,963

Unmanned vehicle for servicing equipment
Guy M Cohen, Lior Horesh, Raya Horesh, Jonathan Lenchner
US Patent App. 15/810,999

Tunable miniaturized physical subsurface model for simulation and inversion
Claudius Feger, Lior Horesh, Ulisses T Mello, Maria JB Moura
US Patent App. 10/108,762

Hamiltonian Engineering with Constrained Optimization for Quantum Sensing and Control
Michael F. O'Keeffe, Lior Horesh, Danielle A. Braje, Isaac L. Chuang
https://arxiv.org/abs/1811.02124 pp. arXiv:1811.02124, 2018

Stable Tensor Neural Networks for Rapid Deep Learning
Elizabeth Newman, Lior Horesh, Haim Avron, Misha Kilmer
https://arxiv.org/abs/1811.06569 pp. arXiv: 1811.06569, 2018

Ultra-sonic self-cleaning system
Guy M Cohen, Lior Horesh, Raya Horesh, Theodore G Van Kessel, Robert L Wisnieff
US Patent App. 15/593,605

Ultra-sonic self-cleaning system
Guy M Cohen, Lior Horesh, Raya Horesh, Theodore G Van Kessel, Robert L Wisnieff
US Patent App. 15/806,698

Optimal experimental design based on mutual information and submodularity
Haim Avron, Guy M Cohen, Lior Horesh, Raya Horesh, Gal Shulkind
US Patent App. 15/614,703

2017

Pseudorandomness and Fourier-Growth Bounds for Width-3 Branching Programs
Thomas Steinke, Salil Vadhan, Andrew Wan
Theory of Computing 13(12), 1--50, 2017

Make up your mind: The price of online queries in differential privacy
Mark Bun, Thomas Steinke, Jonathan Ullman
Proceedings of the Twenty-Eighth Annual ACM-SIAM Symposium on Discrete Algorithms, pp. 1306--1325, 2017

Exposed! A Survey of Attacks on Private Data
Cynthia Dwork, Adam Smith, Thomas Steinke, Jonathan Ullman
Annual Review of Statistics and Its Application pp. 4, Annual Reviews, 2017

Subgaussian Tail Bounds via Stability Arguments
Thomas Steinke, Jonathan Ullman
arXiv preprint arXiv:1701.03493, 2017

Differential Privacy: A Primer for a Non-technical Audience
Kobbi Nissim, Thomas Steinke, Alexandra Wood, Mark Bun, Marco Gaboardi, David R O’Brien, Salil Vadhan
2017

Tight Lower Bounds for Differentially Private Selection
Thomas Steinke, Jonathan Ullman
IEEE Symposium on Foundations of Computer Science (FOCS), 2017

Generalization for Adaptively-chosen Estimators via Stable Median
Vitaly Feldman, Thomas Steinke
Conference on Learning Theory, 2017

Bridging the gap between computer science and legal approaches to privacy
Kobbi Nissim, Aaron Bembenek, Alexandra Wood, Mark Bun, Marco Gaboardi, Urs Gasser, David O’Brien, Thomas Steinke, Salil Vadhan
Harvard Journal of Law \& Technology31, 2017

Calibrating Noise to Variance in Adaptive Data Analysis
Vitaly Feldman, Thomas Steinke
arXiv preprint arXiv:1712.07196, 2017

On Optimal Portfolios of Dynamic Resource Allocations
Lu, Yingdong and Maguluri, Siva Theja and Squillante, Mark S and Wu, Chai Wah
arXiv preprint arXiv:1702.07781, 2017
Abstract

On Optimal Weighted-Delay Scheduling in Input-Queued Switches
Lu, Yingdong and Maguluri, Siva Theja and Squillante, Mark S and Suk, Tonghoon
arXiv preprint arXiv:1704.02302, 2017
Abstract

Hebbian Learning in a Random Network Captures Selectivity Properties of the Prefrontal Cortex
G. W. Lindsay, M. Rigotti, M. R. Warden, E. K. Miller, S. Fusi
Journal of Neuroscience 37(45), 11021-11036, 2017
Abstract

Neural Mechanisms Underlying the Cognitive Regulation of Value Attribution: Abstraction of Contextual Information in the Monkey Hippocampus
S. Bernardi, J. Munuera, M. Rigotti, S. Fusi, C.D. Salzman
Biological Psychiatry 81(10), S272, Elsevier, 2017
Abstract

The hippocampus as a predictive recurrent neural network
S. Recanatesi, G. Wayne, M. Rigotti
Cosyne: Computational and Systems Neuroscience Conference , 2017
Abstract

Modern investigations on the role of the hippocampal system are typically being pursued along two largely independent venues: a spatial navigation view, positing that the hippocampus is part of a complex dedicated to localization in the environment, and a declarative memory view, focusing on rapid encoding and retrieval of episodic experience. Here we investigate a theoretical proposal that tries to reconcile these apparently contrasting views: that the hippocampus supports a semantic relational network organizing semantically related episodes in the service of sequential planning (Eichenbaum and Cohen, 2014). In particular, we provide an algorithmic grounding to this idea in the form of a Recurrent Neural Network (RNN) whose task is to predict future observations in partially observable environments, based on current observations and the actions taken by an actor module. We tested our model by simulating an agent navigating a linear track and a squared 2d arena, and receiving only partial observations about its location. We hypothesized that, in order to accurately predict future observations, our RNN model had to learn to generate recurrent states that usefully summarize the history of its inputs , which, in a Markovian navigation task, would correspond to representing the current location. Indeed, our predictive training procedure leads to recurrent activations that are reminiscent of the well-known hippocampal physiological observations of place cells, border cells, and head-direction cells. Moreover, we show that these learned representations provide good generalization performance in reinforcement learning navigation tasks in partially observable environments. Our model supports the unifying view that both, place-related activity observed in the hippocampus during spatial navigation, as well as its involvement in episodic memory formation, could be a consequence of its role as a semantic relational network (Eichenbaum and Cohen, 2014), specifically via the computational mechanism of a recurrent system trained with predictive coding

Characterizing human prefrontal cortex representations with fMRI
A. Bhandari, M. Rigotti, C. Gagne, S. Fusi, D. Badre
Society for Neuroscience Meeting (Washington DC), 2017
Abstract

Control of Time-Varying Epidemic-Like Stochastic Processes and Their Mean-Field Limits
Lu, Yingdong and Squillante, Mark S and Wu, Chai Wah
arXiv preprint arXiv:1709.10345, 2017
Abstract

Allocating resources among tasks under uncertainty
Lu, Yingdong and Maguluri, Siva Theja and Squillante, Mark S and Wu, Chai Wah
US Patent App. 15/081,827
Abstract

On the control of density-dependent stochastic population processes with time-varying behavior
Lu, Yingdong and Squillante, Mark and Wu, Chai Wah
arXiv preprint arXiv:1709.07988, 2017
Abstract

Dubuque Smarter Travel Project
Ravada, Chandra and Wu, Chai Wah
Technical Report, 2017
Abstract

On Optimal Portfolios of Dynamic Resource Allocations
Lu, Yingdong and Maguluri, Siva Theja and Squillante, Mark S and Wu, Chai Wah
arXiv preprint arXiv:1702.07781, 2017
Abstract

Mechanism for generating a hybrid halftone using a DMSSA screen
Chandu, Kartheek and Stanich, Mikel J and Wu, Chai Wah and Trager, Barry M
US Patent 9,614,998
Abstract

Halftone calibration mechanism employing an ink model
Chandu, Kartheek and Ernst, Larry and Vantaram, Sreenath Rao and Stanich, Mikel J and Wu, Chai Wah
US Patent 9,621,762
Abstract

Efficient solution of quadratically constrained quadratic subproblems within the MADS algorithm
Nadir Amaioua, Charles Audet, Andrew R. Conn and Sebastien Le Digabel
European Journal of Operational Research, 2017
Complete List of Publications

Wireless Sensor Platform for Cultural Heritage Monitoring and Modeling System
L. J. Klein, S. A. Bermudez, A. G. Schrott, M. Tsukada, P. Dionisi-Vici, L. Kargere, F. Marianno, H. F. Hamann, V. Lopez, M. Leona
Sensors 17(9), 2017

Numerical Continuation of Invariant Solutions of the Complex Ginzburg-Landau Equation
V. Lopez
Foundations of Computational Mathematics (FoCM) Conference, 2017

Simulating Heat and Mass Transfer with Limited Amount of Sensor Data
V. Lopez
In: Quintela P. et al. (eds) Progress in Industrial Mathematics at ECMI 2016. Mathematics in Industry, vol 26., Springer, 2017

Globally convergent system and method for automated model discovery
Haim Avron, Lior Horesh, Leo S. Liberti, David Nahamoo
operator, operand, observational study, mathematics, mathematical optimization, machine learning, binary expression tree, artificial intelligence
Abstract

REDUCED SPACE CLUSTERING REPRESENTATIVES AND ITS APPLICATION TO LONG TERM PREDICTION
Andrew R. Conn, Lior Horesh, Ulisses T. Mello, Essen Gijs Michiel Van, Sippe Geert Douma, Eduardo Antonio Jimenez Arismendi
mathematics, machine learning, long term prediction, data mining, cluster analysis, artificial intelligence
Abstract

Image classification using local tensor singular value decompositions
Elizabeth Newman, Misha Kilmer, Lior Horesh
2017 IEEE 7th International Workshop on Computational Advances in Multi-Sensor Adaptive Processing (CAMSAP), pp. 1-5
Abstract tensor contraction, tensor algebra, tensor, symmetric tensor, singular value decomposition, mathematics, machine learning, mnist database, contextual image classification, artificial neural network, artificial intelligence

Location Oriented Membership Code Device
Guy M. Cohen, Lior Horesh, Raya Horesh, Marco Pistoia
computer science

Automatic Eradication of Bio-Related Contaminants from Handles
Amos Cahan, Guy M. Cohen, Lior Horesh, Raya Horesh
ultraviolet, optoelectronics, materials science, contamination, common emitter, coating
Abstract

USING NATURAL LANGUAGE PROCESSING FOR DETECTION OF INTENDED OR UNEXPECTED APPLICATION BEHAVIOR
Lior Horesh, Raya Horesh, Marco Pistoia, Omer Tripp
natural language processing, graphical user interface, computer science, artificial intelligence
Abstract

SYSTEM, METHOD AND APPARATUS FOR DERIVING ROOT CAUSE FOR SOFTWARE TEST FAILURE
Lior Horesh, Raya Horesh, Marco Pistoia, Omer Tripp
timeline, software, root cause, real time computing, observable, flags register, conditional probability, computer science
Abstract

SYSTEM AND METHOD FOR DISCOVERING AD-HOC COMMUNITIES OVER LARGE-SCALE IMPLICIT NETWORKS BY WAVE RELAXATION
Haim Avron, Lior Horesh, Raya Horesh, Omer Tripp
sorting, prefix, mathematics, mathematical optimization, magnitude, machine learning, initial value problem, artificial intelligence
Abstract

METHODS AND APPARATUS FOR COMPUTING ZONAL FLOW RATES IN RESERVOIR WELLS
Lior Horesh, Ankur Narang, Karthik Visweswariah
zonal flow, mathematical optimization, machine learning, generative model, generative grammar, divergence, computer science, artificial intelligence
Abstract

Systems and methods for accelerating Hessian-free optimization for deep neural networks by implicit preconditioning and sampling
Lior Horesh, Brian ED Kingsbury, Tara N Sainath
US Patent 9,601,109

Globally convergent system and method for automated model discovery
Haim Avron, Lior Horesh, Leo S Liberti, David Nahamoo
US Patent App. 14/755,942

Crowd-based detection of device compromise in enterprise setting
Lior Horesh, Raya Horesh, Marco Pistoia, Omer Tripp
US Patent 20,170,093,899

Reduced space clustering representatives and its application to long term prediction
Andrew R Conn, Lior Horesh, Ulisses T Mello, Gijs Michiel van Essen, Sippe Geert Douma, Eduardo Antonio Jimenez Arismendi
US Patent App. 14/922,637

Methods and apparatus for computing zonal flow rates in reservoir wells
Lior Horesh, Ankur Narang, Karthik Visweswariah
US Patent App. 14/951,904

System and method for discovering ad-hoc communities over large-scale implicit networks by wave relaxation
Haim Avron, Lior Horesh, Raya Horesh, Omer Tripp
US Patent App. 14/954,633

Image classification using local tensor singular value decompositions
Elizabeth Newman, Misha Kilmer, Lior Horesh
IEEE CAMSAP, 2017

Medication Dispenser System with Exact Personal Dosing
Amos Cahan, Guy M Cohen, Lior Horesh, Raya Horesh
US Patent App. 14/979,333

Breaking the 49-Qubit Barrier in the Simulation of Quantum Circuits
Edwin Pednault, John A. Gunnels, Giacomo Nannicini, Lior Horesh, Thomas Magerlein, Edgar Solomonik, Robert Wisnieff
https://arxiv.org/abs/1710.05867, 2017

Globally optimal symbolic regression
V. Austel, S. Dash, O. Gunluk, L. Horesh, L. Liberti, G. Nannicini, B. Schieber
NIPS Interpretable ML Workshop, 2017

Location Oriented Membership Code Device
Guy M Cohen, Lior Horesh, Raya Horesh, Marco Pistoia
US Patent App. 15/090,941

Should You Derive, Or Let the Data Drive? An Optimization Framework for Hybrid First-Principles Data-Driven Modeling
Remi R. Lam, Lior Horesh, Haim Avron, Karen E. Willcox
https://arxiv.org/abs/1711.04374, 2017

Automatic Eradication of Bio-Related Contaminants from Handles
Amos Cahan, Guy M Cohen, Lior Horesh, Raya Horesh
US Patent App. 15/157,922

Protecting content displayed on a mobile device
Guy M Cohen, Lior Horesh, Raya Horesh, Marco Pistoia
US Patent 9,703,990

System, method and apparatus for deriving root cause for software test failure
Lior Horesh, Raya Horesh, Marco Pistoia, Omer Tripp
US Patent App. 15/162,979

Reducing Hazards During Mobile Device Use
Guy M Cohen, Lior Horesh, Raya Horesh, Marco Pistoia
US Patent App. 15/163,294

Using natural language processing for detection of intended or unexpected application behavior
Lior Horesh, Raya Horesh, Marco Pistoia, Omer Tripp
US Patent App. 15/159,526

Protecting content displayed on a mobile device
Guy M Cohen, Lior Horesh, Raya Horesh, Marco Pistoia
US Patent 9,684,804

Systems and methods for accelerating hessian-free optimization for deep neural networks by implicit preconditioning and sampling
Lior Horesh, Brian ED Kingsbury, Tara N Sainath
US Patent App. 15/373,775

2016

Bi-level stochastic approximation for joint optimization of hydroelectric dispatch and spot-market operations
Wollega, Ebisa D and Ghosh, Soumyadip and Squillante, Mark S
Winter Simulation Conference (WSC), 2016, pp. 1769--1780
Abstract

Accelerating splitting algorithms for power grid reliability estimation
Wadman, Wander S and Squillante, Mark S and Ghosh, Soumyadip
Proceedings of the 2016 Winter Simulation Conference, pp. 1757--1768
Abstract

Optimal resource capacity management for stochastic networks
AB Dieker, S Ghosh, MS Squillante
Operations Research, Dec 2016, INFORMS

Between Pure and Approximate Differential Privacy
Thomas Steinke, Jonathan Ullman
Journal of Privacy and Confidentiality, 2016

Concentrated differential privacy: Simplifications, extensions, and lower bounds
Mark Bun, Thomas Steinke
Theory of Cryptography Conference, pp. 635--658, 2016

Algorithmic stability for adaptive data analysis
Raef Bassily, Kobbi Nissim, Adam Smith, Thomas Steinke, Uri Stemmer, Jonathan Ullman
Proceedings of the forty-eighth annual ACM symposium on Theory of Computing, pp. 1046--1059, 2016

Upper and Lower Bounds for Privacy and Adaptivity in Algorithmic Data Analysis
Thomas Alexander Steinke
Ph.D. Thesis, 2016

Managing uncertainty in electricity generation and demand forecasting
K Warren, R Ambrosio, B Chen, YH Fu, S Ghosh, Dzung T. Phan, M Sinn, CH Tian, C Visweswariah
IBM Journal of Research and Development 60(1), 8--1, IBM, 2016

Efficient Estimation in the Tails of Gaussian Copulas
Kalyani Nagaraj, Jie Xu, Raghu Pasupathy, Soumyadip Ghosh
arXiv preprint arXiv:1607.01375, 2016

Predicting and mitigating congestion for an electric power system under load and renewable uncertainty
Dzung T Phan, Soumyadip Ghosh
American Control Conference (ACC) , pp. 6791--6796, 2016

STOCHASTIC OPTIMAL DYNAMIC CONTROL OF GI m/GI m/1 n QUEUES WITH TIME-VARYING WORKLOADS
Lu, Yingdong and Sharma, Mayank and Squillante, Mark S and Zhang, Bo
Probability in the Engineering and Informational Sciences 30(03), 470--491, Cambridge Univ Press, 2016
Abstract

Risk-Based Dynamic Allocation of Computing Resources
Lu, Yingdong and Maguluri, Siva Theja and Squillante, Mark S and Wu, Chai Wah
ACM SIGMETRICS Performance Evaluation Review 44(2), 27--29, ACM, 2016
Abstract

Asymptotic optimality of constant-order policies for lost sales inventory models with large lead times
Goldberg, David A and Katz-Rogozhnikov, Dmitriy A and Lu, Yingdong and Sharma, Mayank and Squillante, Mark S
Mathematics of Operations Research, INFORMS, 2016
Abstract

Delay-Optimal Scheduling for Some Input-Queued Switches
Lu, Yingdong and Maguluri, Siva Theja and Squillante, Mark S and Suk, Tonghoon
ACM SIGMETRICS Performance Evaluation Review 44(2), 45--47, ACM, 2016
Abstract

Why neurons mix: high dimensionality for higher cognition
S. Fusi, E.K. Miller, M. Rigotti
Current Opinion in Neurobiology37, 66-74, 2016
Abstract

Energy-efficient neuromorphic classifiers
D. Marti, M. Rigotti, M. Seok, S. Fusi
Neural Computation, 2016
Abstract

Neurophysiological mechanisms for representing abstract components of mental states
J. Munuera, M. Rigotti, S. Fusi, C. D. Salzman
Society for Neuroscience Meeting (San Diego CA), Society for Neuroscience, 2016
Abstract

Contingencies between stimuli, actions, and reinforcement outcomes often differ depending upon the situation. Serial reversal-learning is a behavioural task that can be used to explore the mechanisms underlying such context- dependent adjustments in action selection and reinforcement anticipation. In serial reversal-learning, two or more stimuli reverse their contingencies multiple times within an experimental session. Upon a reversal, agents may update independently their understanding of contingencies of each stimulus- response-outcome association, or they may learn to grasp that within a block of trials, the set of associations for all stimuli de nes a task set. Knowledge of task sets allow a computationally more e cient strategy of exploiting knowledge that the association of one stimulus has changed to infer that contingencies have also switched for the other stimuli within a task set (i.e. by applying inference, a strategy recently demonstrated to be invoked by monkeys, see Saez et al., 2015). This strategy requires the formation and utilization of neural representations of task sets. We investigated the mechanisms that can underlie the formation of representations of task sets. In each of two contexts, rhesus monkeys learned to associate each of 4 images (conditioned stimuli, CSs) to an operant action (hold or release a button) and an unconditioned stimulus (US - either a drop of juice or its absence). Each context was thereby defined by the set of operant and reinforcement contingencies for the CSs, i.e. the task set. Within each context, we manipulated the temporal statistics of events such that particular trial types tended to occur in temporal proximity to each other, creating two sub- contexts. Switches between sub-contexts did not change contingencies so learning about sub-contexts could not be driven by error signals. Instead, learning about sub-contexts relied on the formation of a link between the current trial and the previous trial, a link that could be formed through a Hebbian mechanism if different interconnected neurons represent the current and most recent trial, respectively. We recorded single neuron activity in the amygdala, OFC, and ACC to determine if neural signals merely reflect a memory trace of previous trials, or if they reflect the formation of neural representations of sub-context and context. In all 3 brain areas, the neural representations reflected the linked sets of trials that form sub-contexts and contexts by virtue of the temporal statistics of events. This could underlie a process of abstraction that creates novel, abstract mental states reflecting task sets, a process vital for flexible cognitive behaviour.

Cortico-hippocampal contribution to the generation of contextual information
S. Bernardi, J. Munuera, M. Rigotti, S. Fusi, C. D. Salzman
Society for Neuroscience Meeting (San Diego CA), Society for Neuroscience, 2016
Abstract

Identical stimuli often produce di erent outcomes (and thereby different emotional responses) depending upon the context in which they are encountered. One mechanism the brain employs to infer chances of reward or punishment and to regulate emotional responses is to abstract ‘cognitive contextual structures’. These contexts can be described as the set of circumstances that help us derive situationally informed meaning from the world. Previous studies have established the importance of the prefrontal cortex (PFC) in rule-guided behavior, and recent studies have shown that the amygdala, a brain structure long thought of as a key mediator of emotion, along with the PFC, also represents abstract contextual information. In particular, Saez et al. (2015) showed that the amygdala, ACC and OFC all represent abstract contexts de ned by task sets; a task set is the set of CS-US contingencies in effect within a block of trials. The formation of a neural representation of task set may be created if neural signals providing a memory of a previous trial converge with neural signals representing the current trial. Thus neural signals providing information about the previous trial may be conceptualized as a providing a vital input to the formation of a representation of task set. We sought to determine whether the hippocampus (HPC) or dorsolateral prefrontal cortex (DLPFC) might provide this input by encoding the episodic memory of the previous trial. Alternatively, HPC and DLPFC might explicitly already represent task sets. We have been obtaining simultaneous electrophysiological recording from neurons in the HPC, the ACC and the DLPFC while two rhesus monkeys utilize knowledge of task sets to select an action that maximizes reward rate upon viewing an image. Using a linear decoder, we have observed that all task-relevant variables, including stimulus identity, context, planned operant action, and expected reinforcement may be decoded in each area. However, context could in principle be decoded if neural signals merely reflect a memory trace of the previous trial, since each trial type appears in only one context. We are therefore determining whether the decoding of context derives from signals that represent memory traces, a function long associated with the HPC. Alternatively, neural representations in these areas could reflect the linked trials within a task set. Assuming that neural representations of context truly reflect task sets, and not merely memory traces, a critical question concerns the extent to which these representations are created in a parallel or serial manner across brain structures that include HPC, DLPFC, ACC, OFC, and the amygdala.

Gersgorin variations IV: A left eigenvector approach
Alan J. Hoffman, Chai Wah Wu
Linear Algebra and its Applications498, 136-144, 2016
Abstract singularity, matrix, mathematics, mathematical proof, mathematical analysis, generalized eigenvector, generalization, eigenvalues and eigenvectors, combinatorics

A Simple Proof of a Generalized Cauchy-Binet Theorem
Alan J. Hoffman, Chai Wah Wu
American Mathematical Monthly 123(9), 2016
Abstract proofs of fermat s little theorem, proof of impossibility, picard lindelof theorem, mathematics, mathematical analysis, factor theorem, elementary proof, danskin s theorem, carlson s theorem, analytic proof

Distributed charging of electrical assets
Soumyadip Ghosh, Dung Phan, Mayank Sharma, Chai Wah Wu, JinJun Xiong
US Patent 9,260,031

Online Updating of Computer Model Output Using Real-time Sensor Data
H. Jiang, X. Deng, V. Lopez, and H. Hamann
Technometrics 58(4), 2016

TUNABLE MINIATURIZED PHYSICAL SUBSURFACE MODEL FOR SIMULATION AND INVERSION
Feger, Claudius and Horesh, Lior and Mello, Ulisses T and Moura, Maria JB
US Patent 20,160,098,498

INDUCED CONTROL EXCITATION FOR ENHANCED RESERVOIR FLOW CHARACTERIZATION
Conn, Andrew R and Horesh, Lior and Jimenez Arismendi, Eduardo Antonio and Mello, Ulisses and Van Essen, Gijs
US Patent 20,160,032,692

Protecting content displayed on a mobile device
Cohen, Guy M and Horesh, Lior and Horesh, Raya and Pistoia, Marco
US Patent 9,443,102
Abstract

VIRTUAL SENSING FOR ADJOINT BASED INCORPORATION OF SUPPLEMENTARY DATA SOURCES IN INVERSION
Conn, Andrew R and Douma, Sippe and Van Essen, Gijs and Horesh, Lior and Arismendi Jimenez, Eduardo Antonio and Mello, Ulisses
US Patent 20,160,356,904
Abstract

PROTECTING CONTENT DISPLAYED ON A MOBILE DEVICE
Cohen, Guy M and Horesh, Lior and Horesh, Raya and Pistoia, Marco
US Patent 20,160,358,014
Abstract

Control localization in networks of dynamical systems connected via a weighted tree
Bapat, Ravindra and Wu, Chai Wah
IEEE Transactions on Systems, Man, and Cybernetics: Systems, IEEE, 2016
Abstract

Halftone mechanism
Chandu, Kartheek and Stanich, Mikel J and Wu, Chai Wah and Trager, Barry M
US Patent 9,363,414
Abstract

Toner Estimation Mechanism
Wu, Chai Wah and Trager, Barry M and Stanich, Mikel J and Chandu, Kartheek
US Patent 9,229,408
Abstract

Ger{\v{s}}gorin variations IV: A left eigenvector approach
Hoffman, Alan J and Wu, Chai Wah
Linear Algebra and its Applications498, 136--144, Elsevier, 2016
Abstract

A Simple Proof of a Generalized Cauchy--Binet Theorem
Hoffman, Alan J and Wu, Chai Wah
American Mathematical Monthly 123(9), 928--930, Mathematical Association of America, 2016
Abstract

Meertens number and its variations
Wu, Chai Wah
arXiv preprint arXiv:1603.08493, 2016
Abstract

Sums of products of binomial coefficients mod 2 and run length transforms of sequences
Wu, Chai Wah
arXiv preprint arXiv:1610.06166, 2016
Abstract

Counting the number of isosceles triangles in rectangular regular grids
Wu, Chai Wah
arXiv preprint arXiv:1605.00180, 2016
Abstract

Graph comparison and synchronization in complex networks
Liu, Hui and Cao, Ming and Wu, Chai Wah
Oscillator Circuits: Frontiers in Design, Analysis and Applications32, 183, IET, 2016

Minimal number of points on a grid forming patterns of blocks
Wu, Chai Wah
arXiv preprint arXiv:1608.07247, 2016
Abstract

Removing Clouds and Recovering Ground Observations in Satellite Image Sequences via Temporally Contiguous Robust Matrix Completion
Jialei Wang, Peder A. Olsen, Andrew R. Conn, and Aurelie C. Lozano
Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 2754-2763, 2016
Complete List of Publications

Simulating Heat and Mass Transfer with Limited Amount of Sensor Data
V. Lopez
19th European Conference on Mathematics for Industry (ECMI 2016)

Input device for preventing password theft by thermal imaging
Guy M. Cohen, Lior Horesh, Mayank Sharma
password, key, input device, engineering, embedded system, controller, computer hardware
Abstract

Induced control excitation for enhanced reservoir flow characterization
Andrew R Conn, Lior Horesh, Eduardo Antonio Jimenez Arismendi, Ulisses Mello, Gijs Van Essen
US Patent App. 14/447,165

Tunable miniaturized physical subsurface model for simulation and inversion
Claudius Feger, Lior Horesh, Ulisses T Mello, Maria JB Moura
US Patent App. 14/505,623

Detection and imaging of subsurface high impedance contrast objects
Eldad Haber, Lior Horesh
US Patent App. 14/971,218

Virtual sensing for adjoint based incorporation of supplementary data sources in inversion
Andrew R. Conn, Sippe Douma, Gijs Van Essen, Lior Horesh, Eduardo Antonio Arismendi Jimenez, Ulisses Mello
US Patent 20,160,356,904

Input device for preventing password theft by thermal imaging
Guy M Cohen, Lior Horesh, Mayank Sharma
US Patent App. 14/739,725

System and Method for Adaptive Hierarchical Meta Level Modelling of SAGD Processes
L. Horesh, A.R. Conn, M. Kormaksson, O. Saliu, T. G. van Kessel
Tech Discl Bulletin, 2016

System and Method for Optimal SAGD Operation Control and Sensitivity Analysis by Symbolic Differentiation of Implicit Predictive Models
L. Horesh, A. R. Conn, M. Kormaksson, O. Saliu, T. G. van Kessel
Tech Discl Bulletin, 2016

Globally optimal Mixed Integer Non-Linear Programming (MINLP) formulation for symbolic regression
Lior Horesh, Leo Liberti, Haim Avron
IBM Technical Report pp. ID 219095, 2016

Globally Optimal MINLP Formulation for Symbolic Regression
Lior Horesh, Leo Liberti, Haim Avron
2016

Mathematics of AI gen_feedback_link(leftr, rightr);

links