Author

Alex Tang

15 papers · 2000–2026

2000peak 2 in 20112026

Frequent co-authors

Papers

showing 1–15 of 15 RSS

IZMIT (Kocaeli), Turkey, Earthquake of August 17, 1999, including Duzce Earthquake of November 12, 1999: Lifeline Performance

Alex K. Tang

Artificial Intelligence · 2000 · cited 36× · OpenAlex ↗ · more like this ·

Prepared by the Earthquake Investigation Committee of the Technical Council of Lifeline Earthquake Engineering of ASCE. On August 17, 1999, the Izmit (Kocaeli) earthquake tore through Kocaeli, Turkey, resulting in more than 15,000 deaths and 27,000 injuries; 200,000 people were left homeless. The electric power system, nearly destroyed, sustained more than US$70 million in damage. Another earthquake on November 12, to the east of Izmit at Duzce, added to the lifeline systems damage, impacting the communications system, water and wastewater system, ports, railway, road, and bridges. This TCLEE Monograph provides information for practitioners to improve the performance of the lifeline systems. Topics include: seismology, geology, and geotechnical issues; electric power; airports; telecommunications; water and wastewater; transportation (highways and bridges); liquid fuel facilities…

Telecommunication Systems’ Performance: Christchurch Earthquakes

Alex Tang, Alexis Kwasinski, John Eidinger +2 more

Computer Networks and Communications · Earthquake Spectra · February 2014 · cited 23× · DOI ↗ · more like this ·

Telecommunication systems generally performed better than other lifelines in the Christchurch‐area earthquake sequence of 2010–2011; however, various service interruptions were a major concern for subscribers. Power disruption was the primary reason for service interruption in Christchurch, as has been similarly observed in many other major earthquakes around the world. Extensive ground failures impacted underground cabling, while Central Offices (COs) sustained minor damage due to strong shaking. Closure of the Central Business District and increased call volumes created additional strain on telecommunication service providers to deal with emergency response. This paper presents the findings of the post‐earthquake lifeline performance investigations of both the landline network and the cellular network. Voice and data services of these networks are examined and commented based on the…

Lifelines Performance of the Mw 8.8 off Shore Biobío, Chile Earthquake

Alex K. Tang, Peter J. Eng, C. Eng +1 more

Computer Networks and Communications · Procedia Engineering · 2011 · cited 16× · DOI ↗ · more like this ·

The paper will provide a summary of all lifelines performance in this strong earthquake. The paper will discuss the damage, emergency response, and recovery of electric power, telecommunication, transportation (roads and bridges), seaports, airports, water and wastewater, and their facilities. The impact as a result of the lifelines service interruption will be discussed; most important is the discussion of lifelines interdependence. Each lifeline will be separated into its own section that presents the failure modes, emergency response, and recovery of service. Each section will end with observations and recommendations.

Amazon SageMaker Autopilot: a white box AutoML solution at scale

Piali Das, Valerio Perrone, Nikita Ivkin +22 more

cs.LG · 16 Dec 2020 · cited 7× · arXiv ↗ · more like this ·

AutoML systems provide a black-box solution to machine learning problems by selecting the right way of processing features, choosing an algorithm and tuning the hyperparameters of the entire pipeline. Although these systems perform well on many datasets, there is still a non-negligible number of datasets for which the one-shot solution produced by each particular system would provide sub-par performance. In this paper, we present Amazon SageMaker Autopilot: a fully managed system providing an automated ML solution that can be modified when needed. Given a tabular dataset and the target column name, Autopilot identifies the problem type, analyzes the data and produces a diverse set of complete ML pipelines including feature preprocessing and ML algorithms, which are tuned to generate a leaderboard of candidate models. In the scenario where the performance is not satisfactory, a data…

Telecommunications Performance—May 2008, Wenchuan, Sichuan Earthquake

Alex K. Tang

Computer Networks and Communications · 24 Jun 2009 · cited 5× · DOI ↗ · more like this ·

The May 12, 2008 devastating earthquake in Sichuan Province, China impacted all lifelines. This paper provides a record of observed damage to the telecommunications systems in the earthquake disaster area in Sichuan as a result of the earthquake. A brief over view of the network, the components, and the standards used will be presented. Conclusions of importance of a secured network, and lessons learned will be discussed.

A Longitudinal Study of Tohoku Telecommunication Network Three Years after the Great East Japan Earthquake and Tsunami

Alex Tang, Alexis Kwasinski, Kent Yu

Computer Networks and Communications · 7 Sep 2017 · cited 2× · DOI ↗ · more like this ·

This paper is the first attempt to document the changes and evaluation of the changes to the telecommunication network in Tohoku with respect to improvement and preparedness of future earthquake and tsunami disasters. The information provided in this paper is based on data collected three years after the earthquake and tsunami. A summary of selected locations of significant damage to telecommunication network nodes will be presented. The recovery and rebuilt of these sites will be detailed based on the information collected and followed by evaluation of their impact to resilience with respect to future extreme events. The process and difficulty of collecting recovery and rebuilding information and the primary objective(s) of the service provider during the post disaster environment will also be discussed. This may help to develop guidelines for other lifelines to execute longitudinal…

Tabular Data to Image Generation: Benchmark Data, Approaches, and Evaluation

Alex Tang, Gromit Chan, Ryan A. Rossi +2 more

Computer Vision and Pattern Recognition · 15 Dec 2023 · cited 1× · DOI ↗ · more like this ·

In this work, we study the problem of generating a set of images from an arbitrary tabular dataset. The set of generated images provides an intuitive visual summary of the tabular data that can be quickly and easily communicated and understood by the user. More specifically, we formally introduce this new dataset to image generation task and discuss a few motivating applications including exploratory data analysis and understanding customer segments for creating better marketing campaigns. We then curate a benchmark dataset for training such models, which we release publicly for others to use and develop new models for other important applications of interest. Further, we describe a general and flexible framework that serves as a fundamental basis for studying and developing models for this new task of generating images from tabular data. From the framework, we propose a few different…

Efficient Algorithms for Learning Depth-2 Neural Networks with General ReLU Activations

Pranjal Awasthi, Alex Tang, Aravindan Vijayaraghavan

cs.LG · July 2021 · cited 1× · arXiv ↗ · more like this ·

We present polynomial time and sample efficient algorithms for learning an unknown depth-2 feedforward neural network with general ReLU activations, under mild non-degeneracy assumptions. In particular, we consider learning an unknown network of the form $f(x) = {a}^{\mathsf{T}}σ({W}^\mathsf{T}x+b)$, where $x$ is drawn from the Gaussian distribution, and $σ(t) := \max(t,0)$ is the ReLU activation. Prior works for learning networks with ReLU activations assume that the bias $b$ is zero. In order to deal with the presence of the bias terms, our proposed algorithm consists of robustly decomposing multiple higher order tensors arising from the Hermite expansion of the function $f(x)$. Using these ideas we also establish identifiability of the network parameters under minimal assumptions.

Algorithmic Learning Foundations for Common Law

Jason D. Hartline, Daniel W. Linna Jr., Liren Shan +1 more

cs.CY · 8 Sep 2022 · cited 0× · arXiv ↗ · more like this ·

This paper looks at a common law legal system as a learning algorithm, models specific features of legal proceedings, and asks whether this system learns efficiently. A particular feature of our model is explicitly viewing various aspects of court proceedings as learning algorithms. This viewpoint enables directly pointing out that when the costs of going to court are not commensurate with the benefits of going to court, there is a failure of learning and inaccurate outcomes will persist in cases that settle. Specifically, cases are brought to court at an insufficient rate. On the other hand, when individuals can be compelled or incentivized to bring their cases to court, the system can learn and inaccuracy vanishes over time.

Agnostic Learning of General ReLU Activation Using Gradient Descent

Pranjal Awasthi, Alex Tang, Aravindan Vijayaraghavan

cs.LG · August 2022 · cited 0× · arXiv ↗ · more like this ·

We provide a convergence analysis of gradient descent for the problem of agnostically learning a single ReLU function with moderate bias under Gaussian distributions. Unlike prior work that studies the setting of zero bias, we consider the more challenging scenario when the bias of the ReLU function is non-zero. Our main result establishes that starting from random initialization, in a polynomial number of iterations gradient descent outputs, with high probability, a ReLU function that achieves an error that is within a constant factor of the optimal error of the best ReLU function with moderate bias. We also provide finite sample guarantees, and these techniques generalize to a broader class of marginal distributions beyond Gaussians.

PHYSICAL AND FUNCTIONAL PERFORMANCE OF THE TELECOMMUNICATION INFRASTRUCTURE AFTER THE CANTERBURY EARTHQUAKE SEQUENCE

Sonia Giovinazzi, Adnan Al Rais, Rob Ruiter +5 more

Artificial Intelligence · University of Canterbury Research Repository (University of Canterbury) · 2016 · cited 0× · DOI ↗ · more like this ·

The operation of telecommunication networks is critical during business as usual times, and becomes most vital in post-disaster scenarios,\nwhen the services are most needed for restoring other critical lifelines, due to inherent interdependencies, and for supporting emergency and\nrelief management tasks. In spite of the recognized critical importance, the assessment of the seismic performance for the telecommunication\ninfrastructure appears to be underrepresented in the literature.\nThe FP6 QuakeCoRE project “Performance of the Telecommunication Network during the Canterbury Earthquake Sequence” will provide\na critical contribution to bridge this gap.\nThanks to an unprecedented collaboration between national and international researchers and highly experienced asset managers from Chorus,\ndata and evidences on the physical and functional…

Parallel Collaborative Filtering for Streaming Data

Muqeet Ali, Christopher C. Johnson, Alex K. Tang

Artificial Intelligence · 2011 · cited 0× · OpenAlex ↗ · more like this ·

We present a distributed stochastic gradient descent algorithm for performing low-rank matrix factorization on streaming data. Low-rank matrix factorization is often used as a technique for collaborative filtering. As opposed to recent algorithms that perform matrix factorization in parallel on a batch of training examples [4], our algorithm operates on a stream of incoming examples. We experimentally compare our algorithm with a state-of-art method for performing low-rank matrix factorization on batch data. 1

Body Brush

Horace Ip, Young Hay, Alex Tang

Human-Computer Interaction · 2 Apr 2008 · cited 0× · OpenAlex ↗ · more like this ·

With the development of an innovative motion capture and analysis system using frontal infra-red illumination, and based on a systematic study of the relations between the human body movement and the visual art language, the Body-Brush turns the human body as a whole into a dynamic brush. The Body-Brush enables humans to interact intuitively with the machine to create a rich variety of visual forms and space.

On the Generalization Properties of Selective State-Space Models for Filtering Tasks for Unknown Systems

Alex Tang, M. Emrullah Ildiz, Batin Kurt +2 more

eess.SY · 26 Apr 2026 · arXiv ↗ · more like this ·

Selective State-Space Models (SSMs) such as Mamba have emerged as an alternative architecture to self-attention based transformers in sequence modeling tasks. Recent works have demonstrated the use of transformers in some filtering and output prediction tasks via in-context learning. In this paper, we analyze whether structured SSMs can work equally well for filtering of unknown systems. In particular, we train the SSM on trajectory samples from a set of systems. At run-time, the SSM is given the outputs of an unknown system from the same set and is expected to predict the next output online. Theoretically, under appropriate assumptions, we derive generalization bounds as to why SSMs succeed in such tasks. Empirically, we demonstrate the performance via several numerical examples. We also discuss the advantages and disadvantages of SSMs versus transformers for this task.

Mammo-FM: Breast-specific foundational model for Integrated Mammographic Diagnosis, Prognosis, and Reporting

Shantanu Ghosh, Vedant Parthesh Joshi, Rayan Syed +14 more

cs.CV · December 2025 · arXiv ↗ · more like this ·

Breast cancer is one of the leading causes of death among women worldwide. We introduce Mammo-FM, the first foundation model specifically for mammography, pretrained on the largest and most diverse dataset to date - 140,677 patients (821,326 mammograms) across four U.S. institutions. Mammo-FM provides a unified foundation for core clinical tasks in breast imaging, including cancer diagnosis, pathology localization, structured report generation, and cancer risk prognosis within a single framework. Its alignment between images and text enables both visual and textual interpretability, improving transparency and clinical auditability, which are essential for real-world adoption. We rigorously evaluate Mammo-FM across diagnosis, prognosis, and report-generation tasks in in- and out-of-distribution datasets. Despite operating on native-resolution mammograms and using only one-third of the…