A transdimensional perspective on dark matter
Abstract: Only about 15% of the matter in the Universe interacts electromagnetically and produces light. Although models have been constructed based on the gravitational interactions of dark matter that makes up the remaining 85%, we have not obtained the data that would characterize its particle nature. Today, large data sets are being collected to solve this cosmic problem. However, the bridge we establish between these data sets and our space of hypotheses, must be robust to statistical noise and biases. Therefore, transdimensional inference tools that can take into account modeling degeneracies are needed, when drawing conclusions about the nature of dark matter. By using an inference framework called “Probabilistic Cataloging” we can model data sets such as strongly lensed galaxies or crowded stellar fields, without biases. Furthermore, when dark matter gravitationally lenses a background galaxy, we can probe how dark matter is distributed and structured at small scales. Hence, using data sets expected from future telescopes such as WFIRST, we hope to obtain a smoking gun signature of the yet-unknown interactions of dark matter.