-
The HCI Aspects of Public Deployment of Research Chatbots: A User Study, Design Recommendations, and Open Challenges
Authors:
Morteza Behrooz,
William Ngan,
Joshua Lane,
Giuliano Morse,
Benjamin Babcock,
Kurt Shuster,
Mojtaba Komeili,
Moya Chen,
Melanie Kambadur,
Y-Lan Boureau,
Jason Weston
Abstract:
Publicly deploying research chatbots is a nuanced topic involving necessary risk-benefit analyses. While there have recently been frequent discussions on whether it is responsible to deploy such models, there has been far less focus on the interaction paradigms and design approaches that the resulting interfaces should adopt, in order to achieve their goals more effectively. We aim to pose, ground…
▽ More
Publicly deploying research chatbots is a nuanced topic involving necessary risk-benefit analyses. While there have recently been frequent discussions on whether it is responsible to deploy such models, there has been far less focus on the interaction paradigms and design approaches that the resulting interfaces should adopt, in order to achieve their goals more effectively. We aim to pose, ground, and attempt to answer HCI questions involved in this scope, by reporting on a mixed-methods user study conducted on a recent research chatbot. We find that abstract anthropomorphic representation for the agent has a significant effect on user's perception, that offering AI explainability may have an impact on feedback rates, and that two (diegetic and extradiegetic) levels of the chat experience should be intentionally designed. We offer design recommendations and areas of further focus for the research community.
△ Less
Submitted 7 June, 2023;
originally announced June 2023.
-
Projected $t$-SNE for batch correction
Authors:
Emanuele Aliverti,
Jeff Tilson,
Dayne Filer,
Benjamin Babcock,
Alejandro Colaneri,
Jennifer Ocasio,
Timothy R. Gershon,
Kirk C. Wilhelmsen,
David B. Dunson
Abstract:
Biomedical research often produces high-dimensional data confounded by batch effects such as systematic experimental variations, different protocols and subject identifiers. Without proper correction, low-dimensional representation of high-dimensional data might encode and reproduce the same systematic variations observed in the original data, and compromise the interpretation of the results. In t…
▽ More
Biomedical research often produces high-dimensional data confounded by batch effects such as systematic experimental variations, different protocols and subject identifiers. Without proper correction, low-dimensional representation of high-dimensional data might encode and reproduce the same systematic variations observed in the original data, and compromise the interpretation of the results. In this article, we propose a novel procedure to remove batch effects from low-dimensional embeddings obtained with t-SNE dimensionality reduction. The proposed methods are based on linear algebra and constrained optimization, leading to efficient algorithms and fast computation in many high-dimensional settings. Results on artificial single-cell transcription profiling data show that the proposed procedure successfully removes multiple batch effects from t-SNE embeddings, while retaining fundamental information on cell types. When applied to single-cell gene expression data to investigate mouse medulloblastoma, the proposed method successfully removes batches related with mice identifiers and the date of the experiment, while preserving clusters of oligodendrocytes, astrocytes, and endothelial cells and microglia, which are expected to lie in the stroma within or adjacent to the tumors.
△ Less
Submitted 28 January, 2020; v1 submitted 15 November, 2019;
originally announced November 2019.
-
Structure and Dynamics of the 13/14 November 2012 Eclipse White-Light Corona
Authors:
Jay M. Pasachoff,
Vojtech Rusin,
Metod Saniga,
Bryce A. Babcock,
Muzhou Lu,
Allen B. Davis,
Ronald F. Dantowitz,
Pavlos Gaintatzis,
John H. Seiradakis,
Aristeidis Voulgaris,
Daniel B. Seaton,
Kazuo Shiota
Abstract:
Continuing our series of observations of the motion and dynamics of the solar corona over the solar-activity cycle, we observed the corona from sites in Queensland, Australia, during the 13 (UT)/14 (local time) November 2012 total solar eclipse. The corona took the low-ellipticity shape typical of solar maximum (flattening index ε = 0.01), showing a change from the composite coronal images that we…
▽ More
Continuing our series of observations of the motion and dynamics of the solar corona over the solar-activity cycle, we observed the corona from sites in Queensland, Australia, during the 13 (UT)/14 (local time) November 2012 total solar eclipse. The corona took the low-ellipticity shape typical of solar maximum (flattening index ε = 0.01), showing a change from the composite coronal images that we had observed and analyzed in this journal and elsewhere for the 2006, 2008, 2009, and 2010 eclipses. After crossing the northeast Australian coast, the rest of the path of totality was over the ocean, so further totality was seen only by shipborne observers. Our results include measurements of velocities of a coronal mass ejection; during the 36 minutes of passage from the Queensland coast to a ship north of New Zealand, we find a speed of 413 km/s, and we analyze its dynamics. We discuss the shapes and positions of several types of coronal features seen on our higher-resolution composite Queensland images of the solar corona, including, many helmet streamers, very faint bright and dark loops at the base of helmet streamers, voids and radially oriented thin streamers. We compare our eclipse observations with a hairy-ball model of the magnetic field, confirming the validity of the prediction, and we relate the eclipse phenomenology seen with the near-simultaneous images from the Atmospheric Imaging Assembly on the NASA's Solar Dynamics Observatory (SDO/AIA), the Extreme Ultraviolet Imager on NASA's Solar Terrestrial Relations Observatory (STEREO/EUVI), ESA/ROB's PROBA2/SWAP, and NRL's LASCO on ESA's SOHO. For example, the southeastern CME is related to the solar flare whose origin we trace with a SWAP series of images.
△ Less
Submitted 2 December, 2014;
originally announced December 2014.
-
Revisiting the spreading and covering numbers
Authors:
Ben Babcock,
Adam Van Tuyl
Abstract:
We revisit the problem of computing the spreading and covering numbers. We show a connection between some of the spreading numbers and the number of non-negative integer 2x2 matrices whose entries sum to d, and we construct an algorithm to compute improved upper bounds for the covering numbers.
We revisit the problem of computing the spreading and covering numbers. We show a connection between some of the spreading numbers and the number of non-negative integer 2x2 matrices whose entries sum to d, and we construct an algorithm to compute improved upper bounds for the covering numbers.
△ Less
Submitted 28 May, 2013; v1 submitted 27 September, 2011;
originally announced September 2011.
-
Charon's radius and density from the combined data sets of the 2005 July 11 occultation
Authors:
M. J. Person,
J. L. Elliot,
A. A. S. Gulbis,
J. M. Pasachoff,
B. A. Babcock,
S. P. Souza,
J. Gangestad
Abstract:
The 2005 July 11 C313.2 stellar occultation by Charon was observed by three separate research groups, including our own, at observatories throughout South America. Here, the published timings from the three data sets have been combined to more accurately determine the mean radius of Charon: 606.0 +/- 1.5 km. Our analysis indicates that a slight oblateness in the body (0.006 +/- 0.003) best match…
▽ More
The 2005 July 11 C313.2 stellar occultation by Charon was observed by three separate research groups, including our own, at observatories throughout South America. Here, the published timings from the three data sets have been combined to more accurately determine the mean radius of Charon: 606.0 +/- 1.5 km. Our analysis indicates that a slight oblateness in the body (0.006 +/- 0.003) best matches the data, with a confidence level of 86%. The oblateness has a pole position angle of 71.4 deg +/- 10.4 deg and is consistent with Charon's pole position angle of 67 deg. Charon's mean radius corresponds to a bulk density of 1.63 +/- 0.07 g/cm3, which is significantly less than Pluto's (1.92 +/- 0.12 g/cm3). This density differential favors an impact formation scenario for the system in which at least one of the impactors was differentiated. Finally, unexplained differences between chord timings measured at Cerro Pachon and the rest of the data set could be indicative of a depression as deep as 7 km on Charon's limb.
△ Less
Submitted 3 February, 2006;
originally announced February 2006.
-
Short-Period Waves That Heat the Corona Detected at the 1999 Eclipse
Authors:
Jay M. Pasachoff,
Bryce A. Babcock,
Kevin D. Russell,
Daniel B. Seaton
Abstract:
As a part of a study of the cause of solar coronal heating, we searched for high-frequency (~1 Hz) intensity oscillations in coronal loops in the [Fe XIV] coronal green line. We summarize results from observations made at the 11 August 1999 total solar eclipse from Ramnicu-Valcea, Romania, through clear skies. We discuss the image reduction and analysis through two simultaneous series of coronal…
▽ More
As a part of a study of the cause of solar coronal heating, we searched for high-frequency (~1 Hz) intensity oscillations in coronal loops in the [Fe XIV] coronal green line. We summarize results from observations made at the 11 August 1999 total solar eclipse from Ramnicu-Valcea, Romania, through clear skies. We discuss the image reduction and analysis through two simultaneous series of coronal CCD images digitized at 10 Hz for a total time of about 140 s. One series of images was taken through a 3.6 A filter isolating the 5303 A [Fe XIV] coronal green line and the other through a 100 A filter in the nearby K-corona continuum. Previous observations, described in Pasachoff et al. (2000), showed no evidence for oscillations in the [Fe XIV] green line at a level great than 2% of coronal intensity. We describe several improvements made over the 1998 eclipse that led to increased image clarity and sensitivity. The corona was brighter in 1999 with the solar maximum, further improving the data. We use Fourier analysis to search in the [Fe XIV] channel for intensity oscillations in loops at the base of the corona. Such oscillations in the 1-Hz range are predicted as a result of density fluctuations from the resonant absorption of MHD waves. The dissipation of a significant amount of mechanical energy from the photosphere into the corona through this mechanism could provide sufficient energy to hear the corona. A Monte-Carlo model of the data suggests the presence of enhanced power, particularly in the 0.75-1.0 Hz range, and we conclude that MHD waves remain a viable method for coronal heating.
△ Less
Submitted 12 February, 2002;
originally announced February 2002.
