L. Lee - Institute of Earth Sciences, Academia Sinica, Taipei, Taiwan

L. Lee
Are you L. Lee?

Claim your profile, edit publications, add additional information:

Contact Details

Name
L. Lee
Affiliation
Institute of Earth Sciences, Academia Sinica, Taipei, Taiwan
Country
Taiwan

Pubs By Year

Pub Categories

 
Computer Science - Computation and Language (20)
 
Computer Science - Learning (10)
 
Physics - Physics and Society (9)
 
Nuclear Experiment (6)
 
Mathematics - Numerical Analysis (6)
 
Physics - Instrumentation and Detectors (4)
 
Computer Science - Computer Vision and Pattern Recognition (3)
 
High Energy Physics - Phenomenology (2)
 
Physics - Accelerator Physics (2)
 
Physics - Mesoscopic Systems and Quantum Hall Effect (2)
 
Computer Science - Neural and Evolutionary Computing (2)
 
Computer Science - Computers and Society (2)
 
Computer Science - Artificial Intelligence (2)
 
Computer Science - Human-Computer Interaction (2)
 
Physics - Atomic Physics (1)
 
Physics - Space Physics (1)
 
Nonlinear Sciences - Exactly Solvable and Integrable Systems (1)
 
Quantum Physics (1)
 
Solar and Stellar Astrophysics (1)
 
Mathematics - Spectral Theory (1)
 
Physics - Optics (1)
 
Mathematical Physics (1)
 
Computer Science - Information Retrieval (1)
 
Computer Science - Robotics (1)
 
Physics - Medical Physics (1)
 
Nuclear Theory (1)
 
Physics - Plasma Physics (1)
 
Computer Science - Sound (1)
 
Mathematics - Combinatorics (1)
 
Mathematics - Mathematical Physics (1)

Publications Authored By L. Lee

Concept drift is a major issue that greatly affects the accuracy and reliability of many real-world applications of machine learning. We argue that to tackle concept drift it is important to develop the capacity to describe and analyze it. We propose tools for this purpose, arguing for the importance of quantitative descriptions of drift in marginal distributions. Read More

Despite progress in visual perception tasks such as image classification and detection, computers still struggle to understand the interdependency of objects in the scene as a whole, e.g., relations between objects or their attributes. Read More

The content of today's social media is becoming more and more rich, increasingly mixing text, images, videos, and audio. It is an intriguing research question to model the interplay between these different modes in attracting user attention and engagement. But in order to pursue this study of multimodal content, we must also account for context: timing effects, community preferences, and social factors (e. Read More

Group discussions are a way for individuals to exchange ideas and arguments in order to reach better decisions than they could on their own. One of the premises of productive discussions is that better solutions will prevail, and that the idea selection process is mediated by the (relative) competence of the individuals involved. However, since people may not know their actual competence on a new task, their behavior is influenced by their self-estimated competence --- that is, their confidence --- which can be misaligned with their actual competence. Read More

Headline generation for spoken content is important since spoken content is difficult to be shown on the screen and browsed by the user. It is a special type of abstractive summarization, for which the summaries are generated word by word from scratch without using any part of the original content. Many deep learning approaches for headline generation from text document have been proposed recently, all requiring huge quantities of training data, which is difficult for spoken document summarization. Read More

When large social-media platforms allow users to easily form and self-organize into interest groups, highly related communities can arise. For example, the Reddit site hosts not just a group called food, but also HealthyFood, foodhacks, foodporn, and cooking, among others. Are these highly related communities created for similar classes of reasons (e. Read More

In meetings where important decisions get made, what items receive more attention may influence the outcome. We examine how different types of rhetorical (de-)emphasis -- including hedges, superlatives, and contrastive conjunctions -- correlate with what gets revisited later, controlling for item frequency and speaker. Our data consists of transcripts of recurring meetings of the Federal Reserve's Open Market Committee (FOMC), where important aspects of U. Read More

We have performed a novel comparison between electron-beam polarimeters based on M{\o}ller and Compton scattering. A sequence of electron-beam polarization measurements were performed at low beam currents ($<$ 5 $\mu$A) during the $Q_{\rm weak}$ experiment in Hall C at Jefferson Lab. These low current measurements were bracketed by the regular high current (180 $\mu$A) operation of the Compton polarimeter. Read More

We present a preconditioner based on spectral projection that is combined with a deflated Krylov subspace method for solving ill conditioned linear systems of equations. Our results show that the proposed algorithm requires many fewer iterations to achieve the convergence criterion for solving an ill conditioned problem than a Krylov subspace solver. In our numerical experiments, the solution obtained by the proposed algorithm is more accurate in terms of the norm of the distance to the exact solution of the linear system of equations. Read More

User-machine interaction is important for spoken content retrieval. For text content retrieval, the user can easily scan through and select on a list of retrieved item. This is impossible for spoken content retrieval, because the retrieved items are difficult to show on screen. Read More

Multimedia or spoken content presents more attractive information than plain text content, but the former is more difficult to display on a screen and be selected by a user. As a result, accessing large collections of the former is much more difficult and time-consuming than the latter for humans. It's therefore highly attractive to develop machines which can automatically understand spoken content and summarize the key information for humans to browse over. Read More

Multimedia or spoken content presents more attractive information than plain text content, but it's more difficult to display on a screen and be selected by a user. As a result, accessing large collections of the former is much more difficult and time-consuming than the latter for humans. It's highly attractive to develop a machine which can automatically understand spoken content and summarize the key information for humans to browse over. Read More

Gender bias is an increasingly important issue in sports journalism. In this work, we propose a language-model-based approach to quantify differences in questions posed to female vs. male athletes, and apply it to tennis post-match interviews. Read More

The vector representations of fixed dimensionality for words (in text) offered by Word2Vec have been shown to be very useful in many application scenarios, in particular due to the semantic information they carry. This paper proposes a parallel version, the Audio Word2Vec. It offers the vector representations of fixed dimensionality for variable-length audio segments. Read More

Computational modeling can provide critical insight into existing and potential new surgical procedures, medical or minimally-invasive treatments for heart failure, one of the leading causes of deaths in the world that has reached epidemic proportions. In this paper, we present our Abaqus/Standard-based pipeline to create subject-specific left ventricular models. We first review our generic left ventricular model, and then the personalization process based on magnetic resonance images. Read More

This paper presents the first results to combine two theoretically sound methods (spectral projection and multigrid methods) together to attack ill-conditioned linear systems. Our preliminary results show that the proposed algorithm applied to a Krylov subspace method takes much fewer iterations for solving an ill-conditioned problem downloaded from a popular online sparse matrix collection. Read More

We study a class of mathematical and statistical algorithms with the aim of establishing a computer-based framework for fast and reliable automatic abnormality detection on landmark represented image templates. Under this framework, we apply a landmark-based algorithm for finding a group average as an estimator that is said to best represent the common features of the group in study. This algorithm extracts information of momentum at each landmark through the process of template matching. Read More

Changing someone's opinion is arguably one of the most important challenges of social interaction. The underlying process proves difficult to study: it is hard to know how someone's opinions are formed and whether and how someone's views shift. Fortunately, ChangeMyView, an active community on Reddit, provides a platform where users present their own opinions and reasoning, invite others to contest them, and acknowledge when the ensuing discussions change their original views. Read More

In this work we aim to discover high quality speech features and linguistic units directly from unlabeled speech data in a zero resource scenario. The results are evaluated using the metrics and corpora proposed in the Zero Resource Speech Challenge organized at Interspeech 2015. A Multi-layered Acoustic Tokenizer (MAT) was proposed for automatic discovery of multiple sets of acoustic tokens from the given corpus. Read More

Reinforcement learning (RL) is a general and well-known method that a robot can use to learn an optimal control policy to solve a particular task. We would like to build a versatile robot that can learn multiple tasks, but using RL for each of them would be prohibitively expensive in terms of both time and wear-and-tear on the robot. To remedy this problem, we use the Policy Gradient Efficient Lifelong Learning Algorithm (PG-ELLA), an online multi-task RL algorithm that enables the robot to efficiently learn multiple consecutive tasks by sharing knowledge between these tasks to accelerate learning and improve performance. Read More

In this work, we leverage the linear algebraic structure of distributed word representations to automatically extend knowledge bases and allow a machine to learn new facts about the world. Our goal is to extract structured facts from corpora in a simpler manner, without applying classifiers or patterns, and using only the co-occurrence statistics of words. We demonstrate that the linear algebraic structure of word embeddings can be used to reduce data requirements for methods of learning facts. Read More

Most social network analysis works at the level of interactions between users. But the vast growth in size and complexity of social networks enables us to examine interactions at larger scale. In this work we use a dataset of 76M submissions to the social network Reddit, which is organized into distinct sub-communities called subreddits. Read More

In this paper we propose the Structured Deep Neural Network (structured DNN) as a structured and deep learning framework. This approach can learn to find the best structured object (such as a label sequence) given a structured input (such as a vector sequence) by globally considering the mapping relationships between the structures rather than item by item. When automatic speech recognition is viewed as a special case of such a structured learning problem, where we have the acoustic vector sequence as the input and the phoneme label sequence as the output, it becomes possible to comprehensively learn utterance by utterance as a whole, rather than frame by frame. Read More

The Euler-Poincar\'e (EP) equations describe the geodesic motion on the diffeomorphism group. For template matching (template deformation), the Euler-Lagrangian equation, arising from minimizing an energy function, falls into the Euler-Poincar\'e theory and can be recast into the EP equations. By casting the EP equations in the Lagrangian (or characteristics) form, we formulate the equations as a finite dimensional particle system. Read More

We report on the highest precision yet achieved in the measurement of the polarization of a low energy, $\mathcal{O}$(1 GeV), electron beam, accomplished using a new polarimeter based on electron-photon scattering, in Hall~C at Jefferson Lab. A number of technical innovations were necessary, including a novel method for precise control of the laser polarization in a cavity and a novel diamond micro-strip detector which was able to capture most of the spectrum of scattered electrons. The data analysis technique exploited track finding, the high granularity of the detector and its large acceptance. Read More

Techniques for unsupervised discovery of acoustic patterns are getting increasingly attractive, because huge quantities of speech data are becoming available but manual annotations remain hard to acquire. In this paper, we propose an approach for unsupervised discovery of linguistic structure for the target spoken language given raw speech data. This linguistic structure includes two-level (subword-like and word-like) acoustic patterns, the lexicon of word-like patterns in terms of subword-like patterns and the N-gram language model based on word-like patterns. Read More

This paper presents a new approach for unsupervised Spoken Term Detection with spoken queries using multiple sets of acoustic patterns automatically discovered from the target corpus. The different pattern HMM configurations(number of states per model, number of distinct models, number of Gaussians per state)form a three-dimensional model granularity space. Different sets of acoustic patterns automatically discovered on different points properly distributed over this three-dimensional space are complementary to one another, thus can jointly capture the characteristics of the spoken terms. Read More

This paper presents a novel approach for enhancing the multiple sets of acoustic patterns automatically discovered from a given corpus. In a previous work it was proposed that different HMM configurations (number of states per model, number of distinct models) for the acoustic patterns form a two-dimensional space. Multiple sets of acoustic patterns automatically discovered with the HMM configurations properly located on different points over this two-dimensional space were shown to be complementary to one another, jointly capturing the characteristics of the given corpus. Read More

A new method for the determination of the Alfven wave energy generated during magnetic reconnection is introduced and used to analyze the results from two-dimensional MHD simulations. It is found that the regions with strong Alfven wave perturbations almost coincide with that where both magnetic-field lines and flow-stream lines are bent, suggesting that this method is reliable for identifying Alfven waves. The magnetic energy during magnetic reconnection is mainly transformed into the thermal energy. Read More

This paper summarizes the work done by the authors for the Zero Resource Speech Challenge organized in the technical program of Interspeech 2015. The goal of the challenge is to discover linguistic units directly from unlabeled speech data. The Multi-layered Acoustic Tokenizer (MAT) proposed in this work automatically discovers multiple sets of acoustic tokens from the given corpus. Read More

In this paper we propose the Structured Deep Neural Network (Structured DNN) as a structured and deep learning algorithm, learning to find the best structured object (such as a label sequence) given a structured input (such as a vector sequence) by globally considering the mapping relationships between the structure rather than item by item. When automatic speech recognition is viewed as a special case of such a structured learning problem, where we have the acoustic vector sequence as the input and the phoneme label sequence as the output, it becomes possible to comprehensively learned utterance by utterance as a whole, rather than frame by frame. Structured Support Vector Machine (structured SVM) was proposed to perform ASR with structured learning previously, but limited by the linear nature of SVM. Read More

With the popularity of mobile devices, personalized speech recognizer becomes more realizable today and highly attractive. Each mobile device is primarily used by a single user, so it's possible to have a personalized recognizer well matching to the characteristics of individual user. Although acoustic model personalization has been investigated for decades, much less work have been reported on personalizing language model, probably because of the difficulties in collecting enough personalized corpora. Read More

Wearable devices, or "wearables," bring great benefits but also potential risks that could expose users' activities with- out their awareness or consent. In this paper, we report findings from the first large-scale survey conducted to investigate user security and privacy concerns regarding wearables. We surveyed 1,782 Internet users in order to identify risks that are particularly concerning to them; these risks are inspired by the sensor inputs and applications of popular wearable technologies. Read More

Although analyzing user behavior within individual communities is an active and rich research domain, people usually interact with multiple communities both on- and off-line. How do users act in such multi-community environments? Although there are a host of intriguing aspects to this question, it has received much less attention in the research community in comparison to the intra-community case. In this paper, we examine three aspects of multi-community engagement: the sequence of communities that users post to, the language that users employ in those communities, and the feedback that users receive, using longitudinal posting behavior on Reddit as our main data source, and DBLP for auxiliary experiments. Read More

2014Nov
Authors: MOLLER Collaboration, J. Benesch, P. Brindza, R. D. Carlini, J-P. Chen, E. Chudakov, S. Covrig, M. M. Dalton, A. Deur, D. Gaskell, A. Gavalya, J. Gomez, D. W. Higinbotham, C. Keppel, D. Meekins, R. Michaels, B. Moffit, Y. Roblin, R. Suleiman, R. Wines, B. Wojtsekhowski, G. Cates, D. Crabb, D. Day, K. Gnanvo, D. Keller, N. Liyanage, V. V. Nelyubin, H. Nguyen, B. Norum, K. Paschke, V. Sulkosky, J. Zhang, X. Zheng, J. Birchall, P. Blunden, M. T. W. Gericke, W. R. Falk, L. Lee, J. Mammei, S. A. Page, W. T. H. van Oers, K. Dehmelt, A. Deshpande, N. Feege, T. K. Hemmick, K. S. Kumar, T. Kutz, R. Miskimen, M. J. Ramsey-Musolf, S. Riordan, N. Hirlinger Saylor, J. Bessuille, E. Ihloff, J. Kelsey, S. Kowalski, R. Silwal, G. De Cataldo, R. De Leo, D. Di Bari, L. Lagamba, E. NappiV. Bellini, F. Mammoliti, F. Noto, M. L. Sperduto, C. M. Sutera, P. Cole, T. A. Forest, M. Khandekar, D. McNulty, K. Aulenbacher, S. Baunack, F. Maas, V. Tioukine, R. Gilman, K. Myers, R. Ransome, A. Tadepalli, R. Beniniwattha, R. Holmes, P. Souder, D. S. Armstrong, T. D. Averett, W. Deconinck, W. Duvall, A. Lee, M. L. Pitt, J. A. Dunne, D. Dutta, L. El Fassi, F. De Persio, F. Meddi, G. M. Urciuoli, E. Cisbani, C. Fanelli, F. Garibaldi, K. Johnston, N. Simicevic, S. Wells, P. M. King, J. Roche, J. Arrington, P. E. Reimer, G. Franklin, B. Quinn, A. Ahmidouch, S. Danagoulian, O. Glamazdin, R. Pomatsalyuk, R. Mammei, J. W. Martin, T. Holmstrom, J. Erler, Yu. G. Kolomensky, J. Napolitano, K. A. Aniol, W. D. Ramsay, E. Korkmaz, D. T. Spayde, F. Benmokhtar, A. Del Dotto, R. Perrino, S. Barkanova, A. Aleksejevs, J. Singh

The physics case and an experimental overview of the MOLLER (Measurement Of a Lepton Lepton Electroweak Reaction) experiment at the 12 GeV upgraded Jefferson Lab are presented. A highlight of the Fundamental Symmetries subfield of the 2007 NSAC Long Range Plan was the SLAC E158 measurement of the parity-violating asymmetry $A_{PV}$ in polarized electron-electron (M{\o}ller) scattering. The proposed MOLLER experiment will improve on this result by a factor of five, yielding the most precise measurement of the weak mixing angle at low or high energy anticipated over the next decade. Read More

2014Sep
Authors: Qweak Collaboration, T. Allison, M. Anderson, D. Androic, D. S. Armstrong, A. Asaturyan, T. D. Averett, R. Averill, J. Balewski, J. Beaufait, R. S. Beminiwattha, J. Benesch, F. Benmokhtar, J. Bessuille, J. Birchall, E. Bonnell, J. Bowman, P. Brindza, D. B. Brown, R. D. Carlini, G. D. Cates, B. Cavness, G. Clark, J. C. Cornejo, S. Covrig Dusa, M. M. Dalton, C. A. Davis, D. C. Dean, W. Deconinck, J. Diefenbach, K. Dow, J. F. Dowd, J. A. Dunne, D. Dutta, W. S. Duvall, J. R. Echols, M. Elaasar, W. R. Falk, K. D. Finelli, J. M. Finn, D. Gaskell, M. T. W. Gericke, J. Grames, V. M. Gray, K. Grimm, F. Guo, J. Hansknecht, D. J. Harrison, E. Henderson, J. R. Hoskins, E. Ihloff, K. Johnston, D. Jones, M. Jones, R. Jones, M. Kargiantoulakis, J. Kelsey, N. Khan, P. M. King, E. Korkmaz, S. Kowalski, A. Kubera, J. Leacock, J. P. Leckey, A. R. Lee, J. H. Lee, L. Lee, Y. Liang, S. MacEwan, D. Mack, J. A. Magee, R. Mahurin, J. Mammei, J. W. Martin, A. McCreary, M. H. McDonald, M. J. McHugh, P. Medeiros, D. Meekins, J. Mei, R. Michaels, A. Micherdzinska, A. Mkrtchyan, H. Mkrtchyan, N. Morgan, J. Musson, K. E. Mesick, A. Narayan, L. Z. Ndukum, V. Nelyubin, Nuruzzaman, W. T. H. van Oers, A. K. Opper, S. A. Page, J. Pan, K. D. Paschke, S. K. Phillips, M. L. Pitt, M. Poelker, J. F. Rajotte, W. D. Ramsay, W. R. Roberts, J. Roche, P. W. Rose, B. Sawatzky, T. Seva, M. H. Shabestari, R. Silwal, N. Simicevic, G. R. Smith, S. Sobczynski, P. Solvignon, D. T. Spayde, B. Stokes, D. W. Storey, A. Subedi, R. Subedi, R. Suleiman, V. Tadevosyan, W. A. Tobias, V. Tvaskis, E. Urban, B. Waidyawansa, P. Wang, S. P. Wells, S. A. Wood, S. Yang, S. Zhamkochyan, R. B. Zielinski

The Jefferson Lab Q_weak experiment determined the weak charge of the proton by measuring the parity-violating elastic scattering asymmetry of longitudinally polarized electrons from an unpolarized liquid hydrogen target at small momentum transfer. A custom apparatus was designed for this experiment to meet the technical challenges presented by the smallest and most precise ${\vec{e}}$p asymmetry ever measured. Technical milestones were achieved at Jefferson Lab in target power, beam current, beam helicity reversal rate, polarimetry, detected rates, and control of helicity-correlated beam properties. Read More

Finite difference approximation, in addition to Taylor truncation errors, introduces numerical dispersion-and-dissipation errors into numerical solutions of partial differential equations. We analyze a class of finite difference schemes which are designed to minimize these errors (at the expense of formal order of accuracy), and we analyze the interplay between the Taylor truncation errors and the dispersion-and-dissipation errors during mesh refinement. In particular, we study the numerical dispersion relation of the fully discretized non-dispersive transport equation in one and two space dimensions. Read More

In this paper, we investigate the importance of column scaling in relating two signed-graphic representations of the same matroid. We used the Sage Mathematics software to generate many examples of signed-graphic matroids and their signed-graphic representations. Our examples show that column scaling is sometimes necessary in order to transform one signed-graphic representation into another; moreover, there exist many collections of signed-graphic representations that row-reduce to the same standard form. Read More

Consider a person trying to spread an important message on a social network. He/she can spend hours trying to craft the message. Does it actually matter? While there has been extensive prior work looking into predicting popularity of social-media content, the effect of wording per se has rarely been studied since it is often confounded with the popularity of the author and the topic. Read More

The strength with which a statement is made can have a significant impact on the audience. For example, international relations can be strained by how the media in one country describes an event in another; and papers can be rejected because they overstate or understate their findings. It is thus important to understand the effects of statement strength. Read More

We study a class of partial differential equations (PDEs) in the family of the so-called Euler-Poincar\'e differential systems, with the aim of developing a foundation for numerical algorithms of their solutions. This requires particular attention to the mathematical properties of this system when the associated class of elliptic operators possesses non-smooth kernels. By casting the system in its Lagrangian (or characteristics) form, we first formulate a particles system algorithm in free space with homogeneous Dirichlet boundary conditions for the evolving fields. Read More

The Euler-Poincar\'e differential (EPDiff) equations and the shallow water (SW) equations share similar wave characteristics. Using the Hamiltonian structure of the SW equations with flat bottom topography, we establish a connection between the EPDiff equations and the SW equations in one and multi-dimensions. Additionally, we show that the EPDiff equations can be recast in a curl formulation. Read More

2014Mar
Affiliations: 1CAS Key Laboratory of Geospace Environment, School of Earth and Space Sciences, University of Science and Technology of China, Hefei, China, 2Beijing Institute of Tracing and Telecommunications Technology of China, Beijing, China, 3Institute of Earth Sciences, Academia Sinica, Taipei, Taiwan

A scenario is proposed to explain the preferential heating of minor ions and differential streaming velocity between minor ions and protons observed in the solar corona and in the solar wind. It is demonstrated by test particle simulations that minor ions can be nearly fully picked up by intrinsic Alfv\'en-cyclotron waves observed in the solar wind based on the observed wave energy density. Both high frequency ion-cyclotron waves and low frequency Alfv\'en waves play crucial roles in the pickup process. Read More

We present a conservation formulation and a numerical algorithm for the reduced-gravity shallow-water equations on a beta plane, subjected to a constant wind forcing that leads to the formation of double-gyre circulation in a closed ocean basin. The novelty of the paper is that we reformulate the governing equations into a nonlinear hyperbolic conservation law plus source terms. A second-order fractional-step algorithm is used to solve the reformulated equations. Read More

The lifetime of an atom trap is often limited by the presence of residual background gases in the vacuum chamber. This leads to the lifetime being inversely proportional to the pressure. Here we use this dependence to estimate the pressure and to obtain pressure rate-of-rise curves, which are commonly used in vacuum science to evaluate the performance of a system. Read More

A subset of results from the recently completed Jefferson Lab Qweak experiment are reported. This experiment, sensitive to physics beyond the Standard Model, exploits the small parity-violating asymmetry in elastic ep scattering to provide the first determination of the protons weak charge Qweak(p). The experiment employed a 180 uA longitudinally polarized 1. Read More

We numerically compute eigenvalues of the non-self-adjoint Zakharov--Shabat problem in the semiclassical regime. In particular, we compute the eigenvalues for a Gaussian potential and compare the results to the corresponding (formal) WKB approximations used in the approach to the semiclassical or zero-dispersion limit of the focusing nonlinear Schroedinger equation via semiclassical soliton ensembles. This numerical experiment, taken together with recent numerical experiments [17,18], speaks directly to the viability of this approach; in particular, our experiment suggests a value for the rate of convergence of the WKB eigenvalues to the true eigenvalues in the semiclassical limit. Read More

We show over 100-fold enhancement of the exciton oscillator strength as the diameter of an InGaN nanodisk in a GaN nanopillar is reduced from a few micrometers to less than 40 nm, corresponding to the quantum dot limit. The enhancement results from significant strain relaxation in nanodisks less than 100 nm in diameter. Meanwhile, the radiative decay rate is only improved by 10 folds due to strong reduction of the local density of photon states in small nanodisks. Read More

Single photon emission was observed from site-controlled InGaN/GaN quantum dots. The single-photon nature of the emission was verified by the second-order correlation function up to 90 K, the highest temperature to date for site-controlled quantum dots. Micro-photoluminescence study on individual quantum dots showed linearly polarized single exciton emission with a lifetime of a few nanoseconds. Read More

The Qweak experiment has measured the parity-violating asymmetry in polarized e-p elastic scattering at Q^2 = 0.025(GeV/c)^2, employing 145 microamps of 89% longitudinally polarized electrons on a 34.4cm long liquid hydrogen target at Jefferson Lab. Read More