Luca Cazzanti’s Publications

See also my Google Scholar citations

Scalable and Distributed Sea Port Operational Areas Estimation from AIS data, L. M. Millefiori, D. Zissis, L. Cazzanti, G. Arcieri, IEEE Int. Conf. Data Mining, Barcelona, Spain, December 2016.

Seaports are spatial units that do not remain static over time. They are constantly in flux, evolving according to environmental and connectivity patterns both in size and operational capacity. As such any valid decision making regarding port investment and policy making, essentially needs to take into account port evolution over time and space; thus, accurately defining a seaport’s exact location, operational boundaries, capacity, connectivity indicators, environmental impact and overall throughput. In this work, we apply a data driven approach to defining a seaport’s extended area of operation based on data collected though the Automatic Identification System (AIS). Specifically, we present our adaptation of the well-known KDE algorithm to the MapReduce paradigm, and report results on the port of Rotterdam. PDF


@INPROCEEDINGS{MillefioriIEEEICDM2016,
TITLE = {Scalable and Distributed Sea Port Operational Areas Estimation from {AIS} Data},
AUTHOR = {Millefiori, Leonardo M. and Zissis, Dimitris and Cazzanti, Luca and Arcieri, Gianfranco},
BOOKTITLE = {Int. Conf. Data Mining (ICDM), 2016. Proceedings of },
YEAR = {2016},
MONTH = {December},
ORGANIZATION = {IEEE}
}

Automated Port Traffic Statistics: From Raw Data to Visualisation, L. Cazzanti, A. Davoli, L. M. Millefiori, IEEE Int. Conf. Big Data, Washington, DC, USA, December 2016.

We describe how we leveraged best practices in big data processing pipeline design and visual analytics to prototype the Maritime Patterns-of-Life Information Service (MPoLIS), an information product currently under development at the NATO Centre for Maritime Research and Experimentation (CMRE). MPoLIS supports the maritime industry, governments, and international organizations with visual analytics on vessel traffic in seaports. It addresses three main requirements: a) storing and processing large amounts of data; b) on-demand availability of statistical summaries of vessel traffic in ports; c) intuitive and interactive interface for subject matter experts (SMEs) in the maritime domain. MPoLIS has contributed to building a data-driven, self-service analytics culture within NATO and has been sanctioned for use in support of maritime situational awareness (MSA) in ongoing NATO operations. PDF


@INPROCEEDINGS{CazzantiIEEEBigData2016,
TITLE = {Automated Port Traffic Statistics: From Raw Data to Visualisation},
AUTHOR = {Cazzanti, Luca and Davoli, Antonio and Millefiori, Leonardo M.},
BOOKTITLE = {Big Data, 2016. Proceedings of IEEE},
YEAR = {2016},
MONTH = {December},
ORGANIZATION = {IEEE}
}

A Distributed Approach to Estimating Sea Port Operational Regions from Lots of AIS Data, L. M. Millefiori, D. Zissis, L. Cazzanti, G. Arcieri, IEEE Int. Conf. Big Data, Washington, DC, USA, December 2016.

Seaports play a vital role in the global economy, as they operate as the connection corridors to all other modes of transport and as engines of growth for the wider region. But ports today are faced with numerous unique challenges and for them to remain competitive, significant investments are required. In support of greater transparency in policy making, decisions regarding investment need to be supported by data-driven intelligence. It is often an overlooked fact that seaports do not remain static over time; such spatial units often evolve according to environmental patterns both in size but also connectivity and operational capacity. As such any valid decision making regarding port investment and policy making, essentially needs to take into account port evolution over time and space. In this work, we leverage the huge amounts of vessel data that are progressively becoming available through the Automatic Identification System (AIS) and distributed machine learning to define a seaport’s extended area of operation. Specifically, we present our adaptation of the well-known KDE algorithm to the map-reduce paradigm, and report results on the port of Shanghai. PDF


@INPROCEEDINGS{MillefioriIEEEBigData2016,
TITLE = {A Distributed Approach to Estimating Sea Port Operational Regions from Lots of AIS Data},
AUTHOR = {Millefiori, Leonardo M. and Zissis, Dimitris and Cazzanti, Luca and Arcieri, Gianfranco},
BOOKTITLE = {Big Data, 2016. Proceedings of IEEE},
YEAR = {2016},
MONTH = {December},
ORGANIZATION = {IEEE}
}

Decision Tree-Based Adaptive Modulation for Underwater Acoustic Communications, K. Pelekanakis, L. Cazzanti, G. Zappa, J. Alves, IEEE UCOMMS, Lerici, Italy, September 2016.

Underwater acoustic channels are characterised by non-stationary fading statistics and consequently, a modulation scheme optimally designed for a specific fading model will underperform when the channel statistics change. This issue can be alleviated by using adaptive modulation, i.e., the matching of the modulation scheme to the conditions of the acoustic link. However, selecting signals from a broad range of bit rates is tedious because one needs to know the relationship between the bit error rate (BER) and all relevant channel characteristics (e.g., multipath spread, Doppler spread and signal-to-noise ratio). In this work, this BER-channel relationship is extracted from large amounts of transmissions of a phase-shift keying (PSK) modem. In particular, a decision tree is trained to associate channels with modulation schemes under a target BER. The effectiveness of the proposed tree method is demonstrated by post-processing data from two experimental links off the coast of Faial Island, Azores, Portugal. PDF


@INPROCEEDINGS{PelekanakisUCOMMS2016,
TITLE = {Decision Tree-based Adaptive Modulation for Underwater Acoustic Communications},
AUTHOR = {K. Pelekanakis and L. Cazzanti and G. Zappa and J. Alves},
BOOKTITLE = {Int. Conf. Underwater Communications - {UCOMMS},
PUBLISHER = {{IEEE}},
ADDRESS = {Lerici, Italy},
MONTH = {September},
YEAR = {2016}}

Scalable Estimation of Port Areas from AIS Data, L. M. Millefiori, L. Cazzanti, D. Zissis, G. Arcieri, Workshop on Maritime Knowledge Discovery and Anomaly Detection, Ispra, Italy, July 2016.

This paper discusses work in progress to estimate port locations and operational areas in a scalable, data-driven, unsupervised way. Knowing the extent of port areas is an important component of larger maritime traffic analysis systems that inform stakeholders and decision makers in the maritime industry, governmental agencies, and international organizations. The proposed approach uses Kernel Density Estimator (KDE) and exploits the large volume of Automatic Identification System (AIS) data to learn the extent of port areas in a data-driven way. Example results for the port of La Spezia, Italy, demonstrate the approach for real data. PDF


@INPROCEEDINGS{MillefioriKDAD2016,
TITLE = {Scalable Estimation of Port Areas from {AIS} Data},
AUTHOR = {L. M. Millefiori and L. Cazzanti and D. Zissis and G. Arcieri},
BOOKTITLE = {Workshop on Maritime Knowledge Discovery and Anomaly Detection),
PUBLISHER = {Europen Commission Joint Research Centre},
ADDRESS = {Ispra, Italy},
MONTH = {July},
YEAR = {2016}}

Applications of Nanosatellites for Persistent MSA, L. Cazzanti, CMRE Technical Progress Report, June 2016. [NATO Unclassified]

Maritime Situational Awareness Use Cases Enabled by Space-borne Sensors, L. M. Millefiori, G. Vivone, P. Braca, L. Cazzanti, K. Bryan, SCI-283 Symposium on Considerations for Space and Space-Enabled Capabilities in NATO Coalition Operations, Loghborough, UK, May 2016.

Design of a distributed, fault-tolerant, non-relational database capability for maritime vessel traffic data, L. M. Millefiori, G. arcieri, L. Cazzanti, CMRE Technical Progress Report, April 2016. [NATO Unclassified]

Big Data Architectures in Support of Computational Maritime Situational Awareness - Case Study in Port Traffic Analysis, L. Cazzanti, A. Davoli, CMRE Technical Report, CMRE-FR-2015-021, December 2015. [NATO Unclassified]

A Document-based Data Model in Support of Large Scale Computational Maritime Situational Awareness, L. Cazzanti, L. M. Millefiori, G. Arcieri, IEEE Int. Conf. Big Data, Santa Clara, October 2015.

Computational Maritime Situational Awareness (MSA) supports the maritime industry, governments, and international organizations with machine learning and big data techniques for analyzing vessel traffic data available through the Automatic Identification System (AIS). A critical challenge of scaling computational MSA to big data regimes is integrating the core learning algorithms with big data storage modes and data models. To address this challenge, we report results from our experimentation with MongoDB, a NoSQL document-based database which we test as a supporting platform for computational MSA. We experiment with a document model that avoids database joins when linking position and voyage AIS vessel information and allows tuning the database index and document sizes in response to the AIS data rate. We report results for the AIS data ingested and analyzed daily at the NATO Centre for Maritime Research and Experimentation (CMRE). PDF


@INPROCEEDINGS{@InProceedings{CazzantiIEEEBigData2015,
TITLE = {A Document-based Data Model for Large Scale Computational Maritime Situational Awareness},
AUTHOR = {Cazzanti, Luca and Millefiori, Leonardo M. and Arcieri, Gianfranco},
BOOKTITLE = {Big Data, 2015. Proceedings of IEEE},
YEAR = {2015},
MONTH = {October},
ORGANIZATION = {IEEE}
}

An Information Game to Support the Development of a NATO Harbour Protection Capability, L. Cazzanti, J. Locke, CMRE Memorandum Report, CMRE-2015-MR-007, June 2015. [NATO Unclassified]

Mining Maritime Vessel Traffic: Promises, Challenges, Techniques, L. Cazzanti, G. Pallotta, IEEE/MTS OCEANS, Genova, Italy, May 2015.

This paper discusses machine learning and data mining approaches to analyzing maritime vessel traffic based on the Automated Information System (AIS). We review recent efforts to apply machine learning techniques to AIS data and put them in the context of the challenges posed by the need for both algorithmic performance generalization and interpretability of the results in real-world Maritime Situational Awareness (MSA) settings. We also present preliminary work on discovering and characterizing vessel stationary areas using an unsupervised spatial clustering algorithm. PDF


@INPROCEEDINGS{CazzantiOCEANS2015,
TITLE = {Mining Maritime Vessel Traffic: Promises, Challenges, Techniques},
AUTHOR = {L. Cazzanti and G. Pallotta},
BOOKTITLE = {{OCEANS}},
PUBLISHER = {{IEEE/MTS}},
ADDRESS = {Genova, Italy},
MONTH = {May},
YEAR = {2015}}

Optimization of Surveillance Vessel Network Planning in Maritime Command and Control Systems by Fusing METOC and AIS Vessel Traffic Information, T. Fabbri, R. Vicen-Bueno, R. Grasso, G. Pallotta, L. M. Millefiori, L. Cazzanti, IEEE/MTS OCEANS, Genova, Italy, May 2015.

This paper presents the recent developments of an Optimal Path Planning - Decision Support System (OPP-DSS). The designed framework is based on multi-objective optimization algorithms providing a set of Pareto efficient solutions representing a trade-off among mission objectives. Meteorological and Oceanographic (METOC) and Automatic Identification System (AIS) vessel traffic data are integrated and exploited inside the planning process to improve surveillance in piracy risk areas. Tests in an operational scenario with real-world indication of the effectiveness of the approach. PDF


@INPROCEEDINGS{FabbriOCEANS2015,
TITLE = {Optimization of Surveillance Vessel Network Planning
in Maritime Command and Control Systems by Fusing METOC and AIS Vessel Traffic Information},
AUTHOR = {T. Fabbri and R. Vicen-Bueno and R. Grasso and G. Pallotta and L. M. Millefiori and L. Cazzanti},
BOOKTITLE = {{OCEANS}),
PUBLISHER = {{IEEE/MTS}},
ADDRESS = {Genova, Italy},
MONTH = {May},
YEAR = {2015}}

Grace: A Platform for Optimized 1:1 Marketing at Scale, J. Hersch, S. Miller, L. Cazzanti, B. Flynn, O. Downs, KDD 2015 (unpublished companion paper to invited talk presented by Chief Scientist Olly Downs)

Grace is a software platform for digital marketing that automates the process of empirical evaluation and optimization of marketing campaigns targeting hundreds of millions of customers with thousands of messages. It measures the lift in performance produced by various messages using controlled experiments, and it uses a novel combination of learned decision trees and multi-armed bandits to target each customer with the right message to maximize lift. The exploration/exploitation trade-off is managed automatically by employing a Bayesian approach, Thompson sampling. Performance of the platform is demonstrated with a simulated example, and with a deployed implementation within a prepaid telecom company with millions of subscribers, where Grace generated tangible business impact by increasing revenue compared to business as usual. PDF


@UNPUBLISHED{HershEtAlKDD2025,
TITLE = {Grace: A Platform for Optimized 1:1 Marketing at Scale},
AUTHOR = {J. Hersch, S. Miller, L. Cazzanti, B. Flynn and O. Downs},
NOTE={Companion paper to invited talk presented by Chief Scientist Olly Downs}
ADDRESS = {Seattle, USA},
YEAR = {2015}}

Unsupervised Ranking and Characterization of Differentiated Clusters, L. Cazzanti, C. Mehanian, J. Penzotti, D. Scott, O. Downs, IEEE Conference on Information and Security Informatics, Seattle, June 2013.

We describe a framework for automatically identifying and visualizing the most differentiating attributes of each cluster in a clustered data set. A dissimilarity function measures the cluster-conditional distinguishing saliency of each attribute with respect to a reference realization of the same attribute. For each cluster, the N attributes that are most dissimilar are presented first to the human expert, along with the overall dissimilarity of the cluster. We discuss the computational benefits of the proposed framework, how it can be implemented with map-reduce, its application to the behavioral analysis of mobile phone users, and its broad applicability to diverse problem domains. PDF


@INPROCEEDINGS{CazzantiMehanianEtAlISI2013,
TITLE = {Unsupervised Ranking and Characterization of Differentiated Clusters},
AUTHOR = {L. Cazzanti, C. Mehanian, J. Penzotti, D. Scott and O. Downs},
BOOKTITLE = {Proc. {IEEE} Intelligence and Security Informatics ({ISI}),
ADDRESS = {Seattle, USA},
MONTH = {June},
YEAR = {2013}}

Multi-task Output Regularization, S. Feldman, B. A. Frigyk, M. R. Gupta, L. Cazzanti, P. Sadowski, arXiv:1107.4390v3 [stat.ML], July 2011.

We investigate multi-task learning from an output space regularization perspective. Most multi-task approaches tie together related tasks by constraining them to share input spaces and function classes. In contrast to this, we propose a multi-task paradigm which we call output space regularization, in which the only constraint is that the output spaces of the multiple tasks are related. We focus on a specific instance of output space regularization, multi-task averaging, that is both widely applicable and amenable to analysis. The multi-task averaging estimator improves on the single-task sample average under certain conditions, which we detail. Our analysis shows that for a simple case the optimal similarity depends on the ratio of the task variance to the task differences, but that for more complicated cases the optimal similarity behaves non-linearly. Further, we show that the estimates produced are a convex combination of the tasks’ sample averages. We discuss the Bayesian viewpoint. Three applications of multi-task output space regularization are presented: multi-task kernel density estimation, multi-task-regularized empirical moment constraints in similarity discriminant analysis, and multi-task local linear regression. Experiments on real data sets show statistically significant gains. PDF | Go to the arXiv page.


@ARTICLE{FeldmanFrigykGuptaCazzantiSadowskiarXiv2011,
TITLE = {Multi-task Output Space Regularization},
AUTHOR = {S. Feldman and B. A. Frigyk and M. R. Gupta L. Cazzanti and P. Sadowski},
JOURNAL = {{arXiv}},
EPRINT = {\tt arXiv:1107.4390v3 [stat.ML]}}

Multi-task regularization of generative similarity models, L. Cazzanti. S. Feldman, M. R. Gupta, M. Gabbay, In M. Pelillo and E.R. Hancock (Eds.): SIMBAD 2011, LNCS 7005, pp. 90–103. Springer, Heidelberg (2011).

We investigate a multi-task approach to similarity discriminant analysis, where we propose treating the estimation of the different class-conditional distributions of the pairwise similarities as multiple tasks. We show that regularizing these estimates together using a least-squares regularization weighted by a task-relatedness matrix can reduce the resulting maximum a posteriori classification errors. Results are given for benchmark data sets spanning a range of applications. In addition, we present a new application of similarity-based learning to analyzing the rhetoric of multiple insurgent groups in Iraq. We show how to produce the necessary task relatedness information from standard given training data, as well as how to derive task-relatedness information if given side information about the class relatedness. PDF | Video and slides of the talk

Software - Matlab implementation of multi-task local SDA regularization, plus all other flavors of local SDA.


@INPROCEEDINGS{CazzantiFeldmanGuptaGabbaySIMBAD2011,
TITLE = {Multi-task Regularization of Generative Similarity Models},
AUTHOR = {L. Cazzanti and S. Feldman and M. R. Gupta and M. Gabbay},
BOOKTITLE = {Proc. Intl. Workshop on Similarity-based Pattern Recognition
({SIMBAD 11})},
EDITOR = {M. Pelillo and E. R. Hancock},
VOLUME = {LNCS 7005},
ADDRESS = {Venice, Italy},
MONTH = {September},
YEAR = {2011},
PUBLISHER = {Springer}}

Similarity Discriminant Analysis (SDA) Matlab Toolbox, L. Cazzanti, Applied Physics Laboratory - University of Washington, Seattle, 2011.

This is the Matlab toolbox I developed for my Ph.D. research. It is a set of Matlab scripts for similarity discriminant analysis (SDA), including the standard SDA, local SDA, regularized local SDA, mixture SDA, and nnSDA classfiers. This is research-grade code, intended to test ideas while I was working on my Ph.D. research. I have emphasized readability of the source code rather than speed and memory management. It comes with no guarantees, and I hope you will find it useful. As examples of how to run the software, I have included the scripts I used to test the algorithms on benchmark datasets. This software has benefitted from other people’s helpful suggestions and bug-squashing skills. I want to thank in particular Prof. Maya Gupta of the Dept. EE, University of Washington who was the original force behind the maximum entropy-based approach to estimating similarity distributions. For the theory of the SDA framework for similarity-based classification, see my publications.

Downloads offered without warranty - This is research-grade code

Aug 9, 2011 - sda_20110809.zip. Includes all the previous features, plus multi-task regularized local SDA, and a Matlab implementation of pairwise local SDA.
Mar. 08, 2010 - sda_20100308.zip. Includes mex files for 32-bit Linux (10 times faster than plain m-scripts), a README to get you started quickly, and a sample data set.
Nov. 13, 2009 - sda_20091113.zip. Includes the regularized local SDA classifier, which is the state-of-the-art of SDA-type classifiers. Also icludes code for local BDA.
Jan. 29, 2009 - sda_20090129.zip. First release.


@MISC{CazzantiSdaToolbox2011,
author = “L. Cazzanti”,
year = “2011”,
month = “August”
title = “Similarity Discriminant Analysis Toolbox”,
url = “http://www.lucacazzanti.net/”,
institution = “Applied Physics Laboratory - University of Washington, Seattle”}

Differential frequency hopping performance in Doppler-inducing underwater acoustic communication channels, L. Cazzanti, A. K. Das, D. Egnor, G. Edelson, J. Acoust. Soc. Am. Volume 129, Issue 4, pp. 2665-2665 (May 2011).

The relationships between the parameters of Doppler spread-inducing underwater environments and the bit-error performance of differential frequency hopping (DFH) modulation in the underwater acoustic channel are characterized. Wind speed determines the nature of the effect that the water surface imposes on the acoustic DFH wavefoms propagating underwater. Low wind speeds result in an essentially flat, low-absorption sea surface. In this regime, strong surface reflections and little frequency spreading make intersymbol interference (ISI) the dominant effect on the received waveforms. At high wind speeds, the higher density of air bubbles in the surface layer absorbs almost all energy incident on the surface, resulting in no surface reflections reaching the receiver. In this regime, the surface has little effect on the received signal, either in the form of ISI or frequency spreading. The intermediate ranges of wind speed, with a mix of ISI and surface-induced Doppler spread, pose the most challenging conditions. Simulations and at-sea experiments show that recent algorithmic improvements to the receiver make DFH robust to a variety of environmental conditions and that DFH modulation parameters can be easily adapted to a variety of operationally relevant scenarios based on environmental information.


@ARTICLE{CazzantiDFHDoppler2011,
TITLE = {Differential frequency hopping performance in Doppler-inducing underwater acoustic communication channels},
AUTHOR = {L. Cazzanti and A. K. Das and D. Egnor and G. S. Edelson},
JOURNAL = {J. Acoust. Soc. Am.},
VOLUME = {125},
NUMBER = {4},
PAGES = {2665},
MONTH = {April},
YEAR = {2011}}

Improved Multipath Robustness of DFH Modulation in the Underwater Acoustic Channel, L. Cazzanti, D. Egnor, G. Edelson, A. Das, Proc. IEEE/MTS OCEANS, Seattle, September 2010.

We characterize the performance improvement of differential frequency hopping modulation (DFH) under two techniques that mitigate the multipath effects of the underwater acoustic channel: blind adaptive equalization and beamforming. We report results on data collected at-sea during the RACE08, SPACE08 and WHOI09 experiments, and show that the biterror rate improves with the application of these two techniques. Future improvements may combine joint single-element, blind equalization and beamforming (multi-element) approaches to leverage their respective strengths. PDF


@INPROCEEDINGS{CazzantiEgnorEdelsonDasOCEANS2010,
TITLE = {Improved Multipath Robustness of {DFH} Modulation in the Underwater Acoustic Channel},
AUTHOR = {L. Cazzanti and D. Egnor and G. S. Edelson and A. Das},
BOOKTITLE = {Proc.{IEEE/MTS}{OCEANS} Conference},
ADDRESS = {Seattle, {USA}},
MONTH = {September},
YEAR = {2010}}

Bayesian and Pairwise Local Similarity Discriminant Analysis, P. Sadowski, L. Cazzanti and M. R. Gupta, Proc. Intl. Workshop on Cognitive Information Processing (CIP), Isola d’Elba, Italy, June 2010.

We investigate three extensions to the generative similarity-based classifier called local similarity discriminant analysis (local SDA): a Bayesian approach to estimating the pmfs based on the assumption that similarities are multinomially distributed and on the Dirichlet prior distribution; a pairwise-similarity formulation of local SDA that accounts for all local pairwise similarities to estimate the pmfs; a combined Bayesian pairwise-similarity approach. We discuss how the proposed extensions afford more modeling flexibility than standard local SDA and less cumbersome model training than previously-published local SDA regularization strategies. Experiments with five benchmark similarity-based classification datasets show th at the increased modeling flexibility and lighter computational burden of the proposed extensions are coupled with the good classification performance of the local SDA classification paradigm. PDF

Datasets - AuralSonar | AmazonBinary | Patrol | Voting | Protein

Software - C++ source code for Bayesian local SDA using Dirichlet priors and pairwise local SDA, plus other classifiers we compared to in the paper. Visit the downloads page for more software.


@INPROCEEDINGS{SadowskiCazzantiGuptaCIP2010,
TITLE = {Bayesian and Pairwise Local Similarity Discriminant Analysis},
AUTHOR = {P. Sadowski and L. Cazzanti and M. R. Gupta},
BOOKTITLE = {Proc.Intl. Workshop on Cognitive Information Processing
({CIP})},
ADDRESS = {Isola d'Elba, Italy},
MONTH = {June},
YEAR = {2010}}

Regularizing the Local Similarity Discriminant Analysis Classifier, L. Cazzanti and M. R. Gupta, Proc. Intl. Conf. on Machine Learning and Applications (ICMLA), Miami Beach, December 2009.

We investigate parameter-based and distribution-based approaches to regularizing the generative, similarity-based classifier called local similarity discriminant analysis classifier (local SDA). We argue that regularizing distributions rather than parameters can both increase the model flexibility and decrease estimation variance while retaining the conceptual underpinnings of the local SDA classifier. Experiments with four benchmark similarity-based classification datasets show that the proposed regularization significantly improves classification performance compared to the local SDA classifier, and the distribution- based approach improves performance more consistently than the parameter-based approaches. Also, regularized local SDA can perform significantly better than similarity-based SVM classifiers, particularly on sparse and highly nonmetric similarities. PDF


@INPROCEEDINGS{CazzantiGuptaICMLA2009,
TITLE = {Regularizing the local similarity discriminant analysis classifier},
AUTHOR = {L. Cazzanti and M. R. Gupta},
BOOKTITLE = {Proceedings of the 2009 Intl. Conf. on Machine Learning and Applications ({ICMLA})},
ADDRESS = {Miami Beach, FL, USA},
MONTH = {December},
YEAR = {2009}}

Using Speech Technology to Enhance Isotope ID and Classification, L. M. D. Owsley, J. J. McLaughlin, L. G. Cazzanti, and S. R. Salaymeh, Proc. IEEE Nuclear Science Symposium, Orlando, FL, October 2009.

Learning from Similarities and Other Pairwise Relationships, L. Cazzanti, M. R. Gupta, and Y. Chen. Presented at the Human, Social, Cultural, Behavioral Modeling Workshop (HSCB Focus 2010), Chantilly, August 2009.

Fusing Similarities and Euclidean Features with Generative Classifiers, L. Cazzanti, M.R. Gupta, and S. Srivastava, Proc. Intl. Conf. on Information Fusion (FUSION), Seattle, July, 2009.

Modeling differential frequency hopping communication in the underwater acoustic channel, L. Cazzanti, J. Hsieh, D. Egnor and G. S. Edelson, J. Acoust. Soc. Am. Volume 125, Issue 4, pp. 2581-2581 (April 2009).

Differential frequency hopping (DFH) is a fast frequency hopping digital modulation scheme with proven multiple-access and jamming robustness properties in typical wireless channels. Characterizing the capabilities of DFH modulation in the more challenging underwater acoustic channel requires careful analyses that rely on both computer simulations and data collected at sea. The Sonar Simulation Toolkit (SST) is used to model challenging underwater environments and simulate the propagation of DFH waveforms in the corresponding underwater acoustic channels. Our simulations provide baseline performance results that can be used to assess the capabilities of DFH modulation and guide future algorithmic improvements to the receiver. In particular, our simulations show that incorporating equalization techniques into the DFH baseline receiver leads to improved decoding performance in challenging environments characterized by long channel impulse responses, which are known to cause inter-symbol interference in the received acoustic bit stream. To ensure their relevance to practical applications, our computer-based SST models are calibrated to the environmental parameters measured during recent at-sea experiments, including the Rescheduled Acoustic Communications Experiment (RACE08), and the corresponding performance is analyzed.


@ARTICLE{CazzantiDFHModeling2009,
TITLE = Modeling differential frequency hopping communication in the underwater acoustic channel},
AUTHOR = {L. Cazzanti and J.Hsieh and D. Egnor and G. S. Edelson},
JOURNAL = {J. Acoust. Soc. Am.},
VOLUME = {125},
NUMBER = {4},
PAGES = {2581},
MONTH = {April},
YEAR = {2009}}

Multichannel combination investigations for differential frequency hopping transmissions in shallow water, D. Egnor, G. S. Edelson, L. Cazzanti, and J. Hsieh,J. Acoust. Soc. Am. Volume 125, Issue 4, pp. 2581-2581 (April 2009).

Underwater acoustic communication requires waveforms that are robust to the signal distortions caused by the acoustic channel. Many waveforms used for this purpose require the transmission of training symbols that span the intersymbol interference to learn and compensate for these channel effects. These waveforms also require exacting coordination between the transmitters to avoid multiple access interference. Differential frequency hopping (DFH) is a fast frequency hopping, digital signaling technology that requires minimal information at the transmitter to communicate in the underwater channel. DFH has the desirable performance features of noninterfering spread spectrum operation, spectral reuse, and fading and interference resistance. This presentation describes the baseline autosynchronizing, single-user DFH decoder for a multiple hydrophone receiver and investigates two processing techniques incorporated for shallow-water multiuser applications: fading mitigation and multiuser interference mitigation, as implemented in conjunction with array processing. We present the performance of the baseline DFH decoder in terms of bit error rate with and without these enhancements on single- and multiuser data collected at sea during the 2008 Surface Processes and Acoustic Communications Experiment (SPACE08).


@ARTICLE{EgnorDFHMultichannel2009,
TITLE = Multichannel combination investigations for differential frequency hopping transmissions in shallow water},
AUTHOR = {D. Egnor and G. S. Edelson and L. Cazzanti and J.Hsieh},
JOURNAL = {J. Acoust. Soc. Am.},
VOLUME = {125},
NUMBER = {4},
PAGES = {2581},
MONTH = {April},
YEAR = {2009}}

Differential Frequency Hopping (DFH) Modulation for Underwater Acoustic Communications and Networking [U], D. Egnor, G. S. Edelson, L. Cazzanti, and J. Hsieh, U.S. Navy Journal of Underwater Acoustics, April 2009.

Similarity-based Classification: Concepts and Algorithms, Y. Chen, E. K. Garcia, M. R. Gupta, A. Rahimi, and L. Cazzanti, Journal of Machine Learning Research (JMLR), March 2009.

Underwater Acoustic Single- and Multi-User Differential Frequency Hopping Communications, D. Egnor, L. Cazzanti, J. Hsieh, and G. S. Edelson, Proc. IEEE/MTS OCEANS, Quebec City, 2008.

Generative Models for Similarity-Based Classification, L. Cazzanti, M. R. Gupta, and A. J. Koppal, Pattern Recognition, vol. 41, no. 7, 2289-2297, 2008.

Patent: “Automatic Identification of Sound Recordings, M. Wells, V. Venkatachalam, L. Cazzanti, K. Cheung, N. Dhillon, and S. Sukittanon, USPTO N. 7328153, Feb 5, 2008.

Generative Models for Similarity-based Classification, L. Cazzanti, Ph.D. dissertation, University of Washington, Seattle, 2007.

Local Similarity Discriminant Analysis, L. Cazzanti and M. R. Gupta, Intl. Conf. Machine Learning (ICML), 2007.

Maximum Entropy Generative Models for Similarity-based Learning, M. R. Gupta, L. Cazzanti and A. Koppal, Proc. IEEE Intl. Symposium on Information Theory, 2007.

Information-theoretic and Set-theoretic Similarity, L. Cazzanti and M. R. Gupta, Proc. IEEE Intl. Symposium on Information Theory, 2006.

Minimum Expected Risk Estimation for Near-neighbor Classification, M. R. Gupta, S. Srivastava and L. Cazzanti, UWEE Tech Report, 2006.

Reverse Engineering the Sound of Jazz, L. Cazzanti, V. Hasbrook, M. R. Gupta, UW EE Technical Report Series 2006-0011.

Minimum Expected Risk Probability Estimates for Nonparametric Neighborhood Classifiers, Maya R. Gupta, Luca Cazzanti, and Santosh Srivastava, Proc. IEEE Workshop on Statistical Signal Processing, 2005.

Quality assessment of low free-energy protein structure predictions, L. Cazzanti, M. R. Gupta, L. Malmstroem, and D. Baker, Proc. IEEE Workshop on Machine Learning for Signal Processing, 2005.

Feature-based Modulation Classification Using Circular Statistics, K. Davidson, L. Cazzanti, J. Pitton. and J. Goldschneider, MILCOM, 2004.

Automatic Identification of Sound Recordings, V. Venkatachalam, L. Cazzanti, N. Dhillon, and M. Wells, IEEE Signal Processing Magazine, March 2004.

Blind Demodulation Toolbox - User’s Guide and Reference Manual, L. Cazzanti, K. Davidson, J. Pitton and J. Goldschneider, Insightful Corporation Technical Report, Seattle, WA, March 2004. (Joint work of Insightful Corp. and Applied Physics Laboratory - University of Washington)

Rapid Prototyping of Blind Demodulation Algorithms, J. Goldschneider, L. Cazzanti, D. Stanford and J. Pitton, MathSoft Research Department Technical Report, August 2000.

Positive Time-Frequency Distributions via Numerical Optimization, J. Pitton and L. Cazzanti, MathSoft Research Department Technical Report N. 90, October 1999.

S+DigComm: User Manual, T. Chauvin, J. Goldschneider, J. Pitton, and L. Cazzanti, MathSoft Research Department Technical Report N. 61, October 1998.

S+SigPro: User Manual, J. Pitton and L. Cazzanti, MathSoft Research Department Technical Report N. 58, October 1998.

Parameter Extraction with Neural Networks, L. Cazzanti, M. Khan, and F. Cerrina, SPIE 23rd Annual Symposium on Microlithography, Santa Clara, California, February 1998.

Lithography Parameter Extraction and Optimization Using Neural Networks, L. Cazzanti, Master’s Thesis, University of Wisconsin-Madison, December 1997.