University of Massachusetts Amherst

About: University of Massachusetts Amherst is a education organization based out in Amherst Center, Massachusetts, United States. It is known for research contribution in the topics: Population & Galaxy. The organization has 37274 authors who have published 83965 publications receiving 3834996 citations. The organization is also known as: UMass Amherst & Massachusetts State College.

Mechanism of burial metamorphism of argillaceous sediment: 1. Mineralogical and chemical evidence

Abstract: A detailed mineralogical and chemical investigation has been made of shale cuttings from a well (Case Western Reserve University Gulf Coast 6) in Oligocene-Miocene sediment of the Gulf Coast of the United States. The 10-µm fractions from the 1,250- to 5,500-m stratigraphic interval were analyzed by x-ray diffraction. Major mineralogic changes with depth take place over the interval 2,000 to 3,700 m, after which no significant changes are detectable. The most abundant mineral, illite/smectite, undergoes a conversion from less than 20 percent to about 80 percent illite layers over this interval, after which the proportion of illite layers remains constant. Over the same interval, calcite decreases from about 20 percent of the rock to almost zero, disappearing from progressively larger size fractions with increasing depth; potassium feldspar (but not albite) decreases to zero; and chlorite appears to increase in amount. Variations in the bulk chemical composition of the shale with depth show only minor changes, except for a marked decrease in CaO concomitant with the decrease in calcite. By contrast, the <0.1-µm fraction (virtually pure illite/smectite) shows a large increase in K2O and Al2O3 and a decrease in SiO2 The atomic proportions closely approximate the reaction smectite + Al+3 + K+ = illite + Si+4. The potassium and aluminum appear to be derived from the decomposition of potassium feldspar (and mica?), and the excess silicon probably forms quartz. We interpret all the major mineralogical and chemical changes as the response of the shale to burial metamorphism and conclude that the shale acted as a closed system for all components except H2O, CaO, Na2O, and CO2. Compositional changes in the shale as a function of metamorphic grade closely parallel compositional changes in shale as a function of geologic age.

Topics over time: a non-Markov continuous-time model of topical trends

Abstract: This paper presents an LDA-style topic model that captures not only the low-dimensional structure of data, but also how the structure changes over time. Unlike other recent work that relies on Markov assumptions or discretization of time, here each topic is associated with a continuous distribution over timestamps, and for each generated document, the mixture distribution over topics is influenced by both word co-occurrences and the document's timestamp. Thus, the meaning of a particular topic can be relied upon as constant, but the topics' occurrence and correlations change significantly over time. We present results on nine months of personal email, 17 years of NIPS research papers and over 200 years of presidential state-of-the-union addresses, showing improved topics, better timestamp prediction, and interpretable trends.

Energy and Policy Considerations for Deep Learning in NLP

Abstract: Recent progress in hardware and methodology for training neural networks has ushered in a new generation of large networks trained on abundant data. These models have obtained notable gains in accuracy across many NLP tasks. However, these accuracy improvements depend on the availability of exceptionally large computational resources that necessitate similarly substantial energy consumption. As a result these models are costly to train and develop, both financially, due to the cost of hardware and electricity or cloud compute time, and environmentally, due to the carbon footprint required to fuel modern tensor processing hardware. In this paper we bring this issue to the attention of NLP researchers by quantifying the approximate financial and environmental costs of training a variety of recently successful neural network models for NLP. Based on these findings, we propose actionable recommendations to reduce costs and improve equity in NLP research and practice.

Social Capital and Community Governance

Abstract: Community governance is the set of small group social interactions that, with market and state, determine economic outcomes. We argue (i) community governance addresses some common market and state failures but typically relies on insider-outsider distinctions that may be morally repugnant and economically costly; (ii) the individual motivations supporting community governance are not captured by either selfishness or altruism; (iii) communities, markets and states are complements, not substitutes; (iv) when poorly designed, markets and states crowd out communities; (v) some distributions of property rights are better than others at fostering community governance; and (vi) communities will probably increase in importance in the future.

Nearly 5000 Distant Early-Type Galaxies in COMBO-17: A Red Sequence and Its Evolution since z ~ 1

Abstract: We present the rest-frame colors and luminosities of ~25,000 mR 24 galaxies in the redshift range 0.2 < z ≤ 1.1 drawn from 0.78 deg2 of the COMBO-17 survey (Classifying Objects by Medium-Band Observations in 17 Filters). We find that the rest-frame color distribution of these galaxies is bimodal at all redshifts out to z ~ 1. This bimodality permits a model-independent definition of red early-type galaxies and blue late-type galaxies at any given redshift. The colors of the blue peak become redder toward the present day, and the number density of blue luminous galaxies has dropped strongly since z ~ 1. Focusing on the red galaxies, we find that they populate a color-magnitude relation. Such red sequences have been identified in galaxy cluster environments, but our data show that such a sequence exists over this redshift range even when averaging over all environments. The mean color of the red galaxy sequence evolves with redshift in a way that is consistent with the aging of an ancient stellar population. The rest-frame B-band luminosity density in red galaxies evolves only mildly with redshift in a Λ-dominated cold dark matter universe. When we account for the change in stellar mass-to-light ratio implied by the redshift evolution in red galaxy colors, the COMBO-17 data indicate an increase in stellar mass on the red sequence by a factor of 2 since z ~ 1. The largest source of uncertainty is large-scale structure, implying that considerably larger surveys are necessary to further refine this result. We explore mechanisms that may drive this evolution in the red galaxy population, finding that both galaxy merging and truncation of star formation in some fraction of the blue star-forming population are required to fully explain the properties of these galaxies.