-
Experimenting with Large Language Models and vector embeddings in NASA SciX
Authors:
Sergi Blanco-Cuaresma,
Ioana Ciucă,
Alberto Accomazzi,
Michael J. Kurtz,
Edwin A. Henneken,
Kelly E. Lockhart,
Felix Grezes,
Thomas Allen,
Golnaz Shapurian,
Carolyn S. Grant,
Donna M. Thompson,
Timothy W. Hostetler,
Matthew R. Templeton,
Shinyi Chen,
Jennifer Koch,
Taylor Jacovich,
Daniel Chivvis,
Fernanda de Macedo Alves,
Jean-Claude Paquin,
Jennifer Bartlett,
Mugdha Polimera,
Stephanie Jarmak
Abstract:
Open-source Large Language Models enable projects such as NASA SciX (i.e., NASA ADS) to think out of the box and try alternative approaches for information retrieval and data augmentation, while respecting data copyright and users' privacy. However, when large language models are directly prompted with questions without any context, they are prone to hallucination. At NASA SciX we have developed a…
▽ More
Open-source Large Language Models enable projects such as NASA SciX (i.e., NASA ADS) to think out of the box and try alternative approaches for information retrieval and data augmentation, while respecting data copyright and users' privacy. However, when large language models are directly prompted with questions without any context, they are prone to hallucination. At NASA SciX we have developed an experiment where we created semantic vectors for our large collection of abstracts and full-text content, and we designed a prompt system to ask questions using contextual chunks from our system. Based on a non-systematic human evaluation, the experiment shows a lower degree of hallucination and better responses when using Retrieval Augmented Generation. Further exploration is required to design new features and data augmentation processes at NASA SciX that leverages this technology while respecting the high level of trust and quality that the project holds.
△ Less
Submitted 21 December, 2023;
originally announced December 2023.
-
Improving astroBERT using Semantic Textual Similarity
Authors:
Felix Grezes,
Thomas Allen,
Sergi Blanco-Cuaresma,
Alberto Accomazzi,
Michael J. Kurtz,
Golnaz Shapurian,
Edwin Henneken,
Carolyn S. Grant,
Donna M. Thompson,
Timothy W. Hostetler,
Matthew R. Templeton,
Kelly E. Lockhart,
Shinyi Chen,
Jennifer Koch,
Taylor Jacovich,
Pavlos Protopapas
Abstract:
The NASA Astrophysics Data System (ADS) is an essential tool for researchers that allows them to explore the astronomy and astrophysics scientific literature, but it has yet to exploit recent advances in natural language processing. At ADASS 2021, we introduced astroBERT, a machine learning language model tailored to the text used in astronomy papers in ADS. In this work we:
- announce the first…
▽ More
The NASA Astrophysics Data System (ADS) is an essential tool for researchers that allows them to explore the astronomy and astrophysics scientific literature, but it has yet to exploit recent advances in natural language processing. At ADASS 2021, we introduced astroBERT, a machine learning language model tailored to the text used in astronomy papers in ADS. In this work we:
- announce the first public release of the astroBERT language model;
- show how astroBERT improves over existing public language models on astrophysics specific tasks;
- and detail how ADS plans to harness the unique structure of scientific papers, the citation graph and citation context, to further improve astroBERT.
△ Less
Submitted 29 November, 2022;
originally announced December 2022.
-
SNR G292.0+1.8: A Remnant of a Low-Mass Progenitor Stripped-Envelope Supernova
Authors:
Tea Temim,
Patrick Slane,
John C. Raymond,
Daniel Patnaude,
Emily Murray,
Parviz Ghavamian,
Mathieu Renzo,
Taylor Jacovich
Abstract:
We present a study of the Galactic supernova remnant (SNR) G292.0+1.8, a classic example of a core-collapse SNR that contains oxygen-rich ejecta, circumstellar material, a rapidly moving pulsar, and a pulsar wind nebula (PWN). We use hydrodynamic simulations of the remnant evolution to show that the SNR reverse shock is interacting with the PWN and has most likely shocked the majority of supernova…
▽ More
We present a study of the Galactic supernova remnant (SNR) G292.0+1.8, a classic example of a core-collapse SNR that contains oxygen-rich ejecta, circumstellar material, a rapidly moving pulsar, and a pulsar wind nebula (PWN). We use hydrodynamic simulations of the remnant evolution to show that the SNR reverse shock is interacting with the PWN and has most likely shocked the majority of supernova ejecta. In our models, such a scenario requires a total ejecta mass of $\lesssim 3\: \rm M_{\odot}$ and implies that there is no significant quantity of cold ejecta in the interior of the reverse shock. In light of these results, we compare the estimated elemental masses and abundance ratios in the reverse-shocked ejecta to nucleosynthesis models and find that they are consistent with a progenitor star with an initial mass of 12-16 $\: \rm M_{\odot}$. We conclude that the progenitor of G292.0+1.8 was likely a relatively low mass star that experienced significant mass loss through a binary interaction and would have produced a stripped-envelope supernova explosion. We also argue that the region known as the "spur" in G292.0+1.8 arises as a result of the pulsar's motion through the supernova ejecta and that its dynamical properties may suggest a line-of-sight component to the pulsar's velocity, leading to a total space velocity of $\sim 600\: \rm km\:s^{-1}$ and implying a significant natal kick. Finally, we discuss binary mass loss scenarios relevant to G292.0+1.8 and their implications for the binary companion properties and future searches.
△ Less
Submitted 3 May, 2022;
originally announced May 2022.
-
A Grid of Core-Collapse Supernova Remnant Models I: The Effect of Wind-Driven Mass-Loss
Authors:
Taylor Jacovich,
Daniel Patnaude,
Pat Slane,
Carles Badenes,
Shiu-Hang Lee,
Shigehiro Nagataki,
Dan Milisavljevic
Abstract:
Massive stars can shed material via steady, line-driven winds, eruptive outflows, or mass-transfer onto a binary companion. In the case of single stars, the mass is deposited by the stellar wind into the nearby environment. After the massive star explodes, the stellar ejecta interact with this circumstellar material (CSM), often-times resulting in bright X-ray line emission from both the shock-hea…
▽ More
Massive stars can shed material via steady, line-driven winds, eruptive outflows, or mass-transfer onto a binary companion. In the case of single stars, the mass is deposited by the stellar wind into the nearby environment. After the massive star explodes, the stellar ejecta interact with this circumstellar material (CSM), often-times resulting in bright X-ray line emission from both the shock-heated CSM and ejecta. The amount of material lost by the progenitor, the mass of ejecta, and its energetics all impact the bulk spectral characteristics of this X-ray emission. Here we present a grid of core-collapse supernova remnant models derived from models for massive stars with zero age main sequence masses of $\sim$ 10 - 30 M$_\odot$ evolved from the pre-main sequence stage with wind-driven mass-loss. Evolution is handled by a multi-stage pipeline of software packages. First, we use mesa (Modules for Experiments in Stellar Astrophysics) to evolve the progenitors from pre-main sequence to iron core collapse. We then use the Supernova Explosion Code (snec) to explode the mesa models, and follow them for the first 100 days following core-collapse. Finally, we couple the snec output, along with the CSM generated from mesa mass-loss rates, into the Cosmic-Ray Hydrodynamics code (ChN) to model the remnant phase to 7000 years post core-collapse. At the end of each stage, we compare our outputs with those found in the literature, and we examine any qualitative and quantitative differences in the bulk properties of the remnants and their spectra based on the initial progenitor mass, as well as mass-loss history.
△ Less
Submitted 22 March, 2021; v1 submitted 14 March, 2021;
originally announced March 2021.
-
Modeling Synchrotron Self-Compton and Klein-Nishina Effects in Gamma-Ray Burst Afterglows
Authors:
Taylor Jacovich,
Paz Beniamini,
Alexander van der Horst
Abstract:
We present a self-consistent way of modeling synchrotron self-Compton (SSC) effects in gamma-ray burst afterglows, with and without approximated Klein-Nishina suppressed scattering. We provide an analytic approximation of our results, so that it can be incorporated into the afterglow modeling code \texttt{boxfit}, which is currently based on pure synchrotron emission. We discuss the changes in spe…
▽ More
We present a self-consistent way of modeling synchrotron self-Compton (SSC) effects in gamma-ray burst afterglows, with and without approximated Klein-Nishina suppressed scattering. We provide an analytic approximation of our results, so that it can be incorporated into the afterglow modeling code \texttt{boxfit}, which is currently based on pure synchrotron emission. We discuss the changes in spectral shape and evolution due to SSC effects, and comment on how these changes affect physical parameters derived from broadband modeling. We show that SSC effects can have a profound impact on the shape of the X-ray light curve using simulations including these effects. This leads to data that cannot be simultaneously fit well in both the X-ray and radio bands when considering synchrotron-only fits, and an inability to recover the correct physical parameters, with some fitted parameters deviating orders of magnitude from the simulated input parameters. This may have a significant impact on the physical parameter distributions based on previous broadband modeling efforts.
△ Less
Submitted 8 July, 2020;
originally announced July 2020.
