Search

Twenty-five hundred years ago, the paperwork of the Achaemenid Empire was recorded on clay tablets. In 1933, archaeologists from the University of Chicago's Oriental Institute (OI) found tens of thousands of these tablets and fragments during the excavation of Persepolis. Many of these tablets have been painstakingly photographed and annotated by expert cuneiformists, and now provide a rich dataset consisting of over 5,000 annotated tablet images and 100,000 cuneiform sign bounding boxes. We leverage this dataset to develop DeepScribe, a modular computer vision pipeline capable of localizing cuneiform signs and providing suggestions for the identity of each sign. We investigate the difficulty of learning subtasks relevant to cuneiform tablet transcription on ground-truth data, finding that a RetinaNet object detector can achieve a localization mAP of 0.78 and a ResNet classifier can achieve a top-5 sign classification accuracy of 0.89. The end-to-end pipeline achieves a top-5 classification accuracy of 0.80. As part of the classification module, DeepScribe groups cuneiform signs into morphological clusters. We consider how this automatic clustering approach differs from the organization of standard, printed sign lists and what we may learn from it. These components, trained individually, are sufficient to produce a system that can analyze photos of cuneiform tablets from the Achaemenid period and provide useful transliteration suggestions to researchers. We evaluate the model's end-to-end performance on locating and classifying signs, providing a roadmap to a linguistically-aware transliteration system, then consider the model's potential utility when applied to other periods of cuneiform writing., Comment: Currently under review in the ACM JOCCH

We present new Spitzer transit observations of four K2 transiting sub-Neptunes: K2-36c, K2-79b, K2-167b, and K2-212b. We derive updated orbital ephemerides and radii for these planets based on a joint analysis of the Spitzer, TESS, and K2 photometry. We use the EVEREST pipeline to provide improved K2 photometry, by detrending instrumental noise and K2's pointing jitter. We used a pixel level decorrelation method on the Spitzer observations to reduce instrumental systematic effects. We modeled the effect of possible blended eclipsing binaries, seeking to validate these planets via the achromaticity of the transits (K2 versus Spitzer). However, we find that Spitzer's signal-to-noise ratio for these small planets is insufficient to validate them via achromaticity. Nevertheless, by jointly fitting radii between K2 and Spitzer observations, we were able to independently confirm the K2 radius measurements. Due to the long time baseline between the K2 and Spitzer observations, we were also able to increase the precision of the orbital periods compared to K2 observations alone. The improvement is a factor of 3 for K2-36c, and more than an order of magnitude for the remaining planets. Considering possible JWST observations in 1/2023, previous 1 sigma uncertainties in transit times for these planets range from 74 to 434 minutes, but we have reduced them to the range of 8 to 23 minutes., Comment: 18 pages, 11 Figures, Accepted by The Astronomical Journal

Robots operating alongside humans in diverse, stochastic environments must be able to accurately interpret natural language commands. These instructions often fall into one of two categories: those that specify a goal condition or target state, and those that specify explicit actions, or how to perform a given task. Recent approaches have used reward functions as a semantic representation of goal-based commands, which allows for the use of a state-of-the-art planner to find a policy for the given task. However, these reward functions cannot be directly used to represent action-oriented commands. We introduce a new hybrid approach, the Deep Recurrent Action-Goal Grounding Network (DRAGGN), for task grounding and execution that handles natural language from either category as input, and generalizes to unseen environments. Our robot-simulation results demonstrate that a system successfully interpreting both goal-oriented and action-oriented task specifications brings us closer to robust natural language understanding for human-robot interaction., Comment: Accepted at the 1st Workshop on Language Grounding for Robotics at ACL 2017

Belief updating in Bayes nets, a well known computationally hard problem, has recently been approximated by several deterministic algorithms, and by various randomized approximation algorithms. Deterministic algorithms usually provide probability bounds, but have an exponential runtime. Some randomized schemes have a polynomial runtime, but provide only probability estimates. We present randomized algorithms that enumerate high-probability partial instantiations, resulting in probability bounds. Some of these algorithms are also sampling algorithms. Specifically, we introduce and evaluate a variant of backward sampling, both as a sampling algorithm and as a randomized enumeration algorithm. We also relax the implicit assumption used by both sampling and accumulation algorithms, that query nodes must be instantiated in all the samples., Comment: Appears in Proceedings of the Twelfth Conference on Uncertainty in Artificial Intelligence (UAI1996)

We perform 1D particle-in-cell (PIC) simulations using OSIRIS, which model a short-duration (~500/{\omega}0 FWHM) scattered light seed pulse in the presence of a constant counter-propagating pump laser with an intensity far below the absolute instability threshold. The seed undergoes linear convective Raman amplification and dominates over fluctuations due to particle discreteness. Our simulation results are in good agreement with results from a coupled mode solver when we take into account special relativity and the use of finite size PIC simulation particles. We present linear gain spectra including both effects. Extending the PIC simulations past when the seed exits the simulation domain reveals bursts of large-amplitude scattering in many cases, which does not occur in simulations without the seed pulse. These bursts can have amplitudes several times greater than the amplified seed pulse, and we demonstrate that this large-amplitude scattering is the result of kinetic inflation by examining trapped particle orbits. This large-amplitude scattering is caused by the seed modifying the distribution function earlier in the simulation. We perform some simulations with longer duration seeds, which lead to parts of the seeds undergoing kinetic inflation and reaching amplitudes several times more than the steady-state linear theory results. Simulations with continuous seeds demonstrate that the onset of inflation depends on seed wavelength and incident intensity, and we observe oscillations in the reflectivity at a frequency equal to the difference between the seed frequency and the frequency at which the inflationary SRS grows., Comment: 17 pages, 23 figures, published in Physics of Plasmas. Copyright 2012 American Institute of Physics

In this paper we identify the ``nil-terms'' for Waldhausen's algebraic K-theory of spaces functor as the reduced K-theory of a category of equivariant spaces equipped with a homotopically nilpotent endomorphism., Comment: 10 pages