Search

Interpretability, transparency, and auditability of machine learning (ML)-driven investment has become a key issue for investment managers as many look to enhance or replace traditional factor-based investing. The authors show that symbolic artificial intelligence (SAI) provides a solution to this conundrum, with superior return characteristics compared to traditional factor-based stock selection, while producing interpretable outcomes. Their SAI approach is a form of satisficing that systematically learns investment decision rules (symbols) for stock selection, using an a priori algorithm, avoiding the need for error-prone approaches for secondary explanations (known as XAI). The authors compare the empirical performance of an SAI approach with a traditional factor-based stock selection approach, in an emerging market equities universe. They show that SAI generates superior return characteristics and would provide a viable and interpretable alternative to factor-based stock selection. Their approach has significant implications for investment managers, providing an ML alternative to factor investing but with interpretable outcomes that could satisfy internal and external stakeholders. Key Findings ▪ Symbolic artificial intelligence (SAI) for stock selection, a form of satisficing, provides an alternative to factor investing and overcomes the interpretability issues of many machine learning (ML) approaches. ▪ An SAI that could be applied at scale is shown to produce superior return characteristics to traditional factor-based stock selection. ▪ SAI’s superior stock selection is examined using notional visualizations of its decision boundaries.

A source mechanism for the noble gases in the earth's atmosphere may be the solar-wind bombardment of extraterrestrial dust. Calculations based on estimates of recent influx of dust to the earth, quiet solar-wind intensity, argon abundance in the solar wind, and assumed Poynting-Robertson lifetimes of small particles in space agree within a factor of 5 with the present amounts of argon-36 and argon-38 in the atmosphere.

Isotopic abundance variations of silicon in quartz and feldspar from the Rose Quartz Pegmatite of the Pala district in Southern California are reported. All feldspar samples are enriched in Si30 with respect to adjacent quartz samples, and the magnitude of enrichment varies from 0.6 to 2.7 per mil in Si30. Feldspar-quartz differences were found to increase from outer zones towards the inner quartz core. All measurements on quartz samples were within a total range of about 0.6 per mil. The feldspars most heavily enriched in Si30 are believed to result from Rayleigh fractionation during crystallization from a vapor phase.

Measurements of excess Ar(36) + Ar(38) ( released mainly at 1200 degrees C) in magnetic concentrates of Pacific sediments and in a dense concentrate of Greenland dust agree within an order of magnitude with expected concentrations implanted by solar-flare ion streams of energy less than 10 Mev per atomic-mass unit. The agreement implies that more than 10 percent of each concentrate may be extraterrestrial, depending on size distribution and flare spectra. Rare-gas measurements on fine-grained dust can provide data on: solar-flare "paleo-ion" fluxes, energy spectra, and isotopic abundances; identification, mineralogy, and chemistry of interplanetary dust; influx rates to Earth and sedimentation rates of oceanic cores; and lunar-surface residence and mixing times.

Measurements of natural variations in isotopic abundance ratios of silicon are reported. A maximum natural range of 5.3 per mil in Si/sup 30/ is obsenved to date. Coexisting biotite, quartz, and feldspars from some specimens of igneous rock from the Yosemite region differ by as much as 3 per mil; in other grossly similar igneous rocks from the same region the same minerals differ by less than 0.3 per mil. In those rocks with large differences between the minerals, the ratio of STASi/sup 30/!/ STASi/sup 28/ + Si/sup 29/! increases in the order biotite, quartz, feldspar. All granitic rocks studied are enriched in Si/sup 30/ with respect to meteorites and basic rocks. Measurements on some sediments and biologic samples covered a range of approximately 4 per mil. A group of tektites covered a range of less than 0.7 per mil, with isotopic abundances centrally distributed within the observed normal terrestrial range. The results are interpreted, and applications of such studies to problems of geothermometry and formation of igneous rocks are discussed. (auth)

RECENT measurements of hydrogen diffusion-rates in α-iron by Johnson and Hill1 gave approximately 2 × 10−7 cm.2/sec. for the value of D at room temperature; and in steel by Frank, Lee, and Williams2 gave approximately 1 × 10−7 cm.2/sec. Johnson and Hill measured re-emission of hydrogen from cylinders previously soaked in hydrogen at high temperatures and pressures, while Frank et al. used a mass spectrometer to measure the emission of hydrogen and deuterium from thin sheets. Frank, Swets and Fry3, observing the build-up of hydrogen permeating through thin sheets of steel, obtained for D at room temperature values about an order of magnitude higher.

IN a recent communication, Tamers1 has suggested that tritium in iron may decay radioactively with a time short compared with its twelve-year half-life, perhaps with a half-life of several days or less. He bases this suggestion principally on some misinterpretations of tritium measurements in iron meteorites. It may be useful to summarize the experimental meteorite evidence, which, if examined in detail, cannot be used to support such a suggestion. I should like here to present some of this evidence and to indicate briefly and in preliminary form a simple, plausible and physically reasonable explanation for the tritium observations in the iron–nickel phases of meteorites. A discussion of the other physical reasons for not believing Tamers's suggestion seems completely superfluous.

SEVERAL points urgently require clarification following my recent exchange of communications with Mr. Tamers1,2. First and most important, the limited available evidence does indeed suggest that a difference of several orders of magnitude exists between γ- and α-phase of iron for diffusion constants of hydrogen at room temperature. It is, of course, the room-temperature behaviour which is relevant for the problem of loss on Earth. Mr. Tamers quotes the paper of Sykes, Burton and Gegg3 as showing that these two diffusion constants differ by only a factor of 10 at 650° C. The identical equations and numbers from this paper demonstrate a difference of a factor of one million at room temperature (20° C).

The present rapid advance in space sciences, clearly apparent at previous semiyearly national meetings of the American Geophysical Union, was again in evidence at the Stanford meeting, December 27–29, 1962. A total of over 60 papers was presented by the Planetary Sciences Section, usually in sessions held jointly with other Sections, including Geomagnetism and Aeronomy; Meteorology; and Volcanology, Geochemistry, and Petrology. Two symposia, jointly sponsored by the American Physical Society, also included items of space interest, such as the atmospheric drag of satellites and the Mariner 2 probe observations on Venus conducted a few days previous to the meeting.