The Overproduction of truth: Passion, competition, and integrity in modern science

The way science is done has changed radically in recent years. Scientific research and institutions, which have long been characterized by passion, dedication and reliability, have increasingly less capacity for more ethical pursuits, and are pressed by hard market laws. From the vocation of a few, science has become the profession of many — possibly too many. These trends come with consequences and risks, such as the rise in fraud, plagiarism, and in particular the sheer volume of scientific publications, often of little relevance. We will critically review and assess the present-day policies and behaviors in scientific production and publication. We will touch on the tumultuous growth of scientific journals, in parallel with the growth of self-declared scientists over the world. We will investigate the loopholes and hoaxes of pretend journals and nonexistent congresses, so common today in the scientific arena, and discuss problems connected to the incorrect use of bibliometric indices, which have resulted in large part from the above distortions of scientific life. The solution? A slow approach with more emphasis on quality rather than quantity that will help us to rediscover the essential role of the responsible scientist.

Reference: G. Pacchioni, The Overproduction of truth. Passion, competition, and integrity in modern science, Oxford University Press, Oxford 2018.

Gianfranco Pacchioni received his PhD at the Freie Universität Berlin in 1984. He worked at the IBM Almaden Research Center, and at the Technical University of Munich. He is Full Professor at the University of Milano Bicocca where he has been Vice Rector for Research (2013-2019) and Director of the Department of Materials Science (2003-2009). He has published more than 500 papers (h index 85; WoS) and given nearly 500 invited talks on the electronic structure of oxide surfaces and interfaces, 2D oxides, defects, supported metal clusters, and catalysis. He received several awards and is Fellow of the Accademia Nazionale dei Lincei (2014), the Academia Europaea (2012), and the European Academy of Sciences (2009). He has authored some popular science books, including “The overproduction of truth”, Oxford University Press, 2018, where he presents a personal view of the problems of contemporary science.

Watch here:
https://umd.zoom.us/rec/share/impshYCSYMhBBqYm4hZx01aCzCEasXE-5emDf4qLDKC6mTNRyx2BErNJP_XfZ5QS.F_6TgBOs4_UvJ1rW

Access Passcode: 7kml7&^P

None scheduled   

 

Collision Course: Particle Physics Meets Machine Learning

Modern machine learning has had an outsized impact on many scientific fields, and particle physics is no exception. What is special about particle physics, though, is the vast amount of theoretical and experimental knowledge that we already have about many problems in the field. In this colloquium, I present two cases studies involving quantum chromodynamics (QCD) at the Large Hadron Collider (LHC), highlighting the fascinating interplay between theoretical principles and machine learning strategies. First, by cataloging the space of all possible QCD measurements, we (re)discovered technology relevant for self-driving cars. Second, by quantifying the similarity between two LHC collisions, we unlocked a class of nonparametric machine learning techniques based on optimal transport. In addition to providing new quantitative insights into QCD, these techniques enable new ways to visualize data from the LHC.

Watch here:
https://umd.zoom.us/rec/share/clp5iCCm4QVBB25M4hNaU505T72f9A82pqkHOmHSdjnfQJ0cBhK3bJaaeOtpx-yy.cBtja9SPD2Ok0sMv

Access Passcode: q=cY394#  

None scheduled  

 

Twistronics in solid-state devices and beyond

The ability to control and manipulate the strength of correlations in quantum matter is one of the central questions in condensed matter physics today. While pressure, chemical doping, or magnetic field have served as conventional tuning knobs for a wide class of correlated systems, the ability to twist van der Waals materials has recently emerged as a novel scheme to engineer strong correlations and tune electronic properties. In the case of twisting two sheets of graphene, at a particular ``magic-angle,’’ the kinetic energy of electronic degrees of freedom is expected to vanish, and as a result, interaction effects should dominate. This has now been demonstrated experimentally following the recent discovery of superconductivity in close proximity to correlated insulating phases in magic-angle graphene. These results have now been reproduced by a number of experimental groups and extended to other two-dimensional van der Waals materials.

In this talk, I will discuss our theory that describes the magic-angle phenomena as a universal property of Dirac points in an incommensurate potential. As a result, we will discuss the generalization of the magic-angle effect to a wide class of models and distinct physical settings, such as ultra-cold atomic gases, trapped ions, and metamaterials. It will be shown that the magic-angle in twisted bilayer graphene is, in fact, a single particle quantum phase transition that can be described by a delocalization transition in momentum space with multifractal wavefunctions akin to electrons in a magnetic field and the celebrated Hofstadter's butterfly. The effects of correlations will be considered by constructing effective Hubbard models with dramatically enhanced interactions due to this novel quantum phase transition.

Quantum Chaos and the Foundations of Statistical Mechanics

The question of how isolated many-body systems come to thermal equilibrium has been debated since before the advent of quantum mechanics, which seems only to confuse the issue further. I will present a simple picture, developed over the last two decades, of how thermalization of isolated quantum many-body systems can occur, with the universal properties of chaotic quantum dynamics as the underlying mechanism. An array of analytic, numerical, and experimental evidence now supports the validity of this picture for a large class of systems.  

ZOOM link: https://umd.zoom.us/j/95368896289?pwd=d0pqN3FGWjdCSFJlZDdXSmcxNlhLZz09 Meeting ID: 953 6889 6289 Passcode: 89593

Where Computational Science Meets Experiment: Self-Driving Laboratories for Materials Discovery

Register for the lecture: https://go.umd.edu/shih-i-pai-20

To face the challenges of the 21st century, modern society requires the rapid discovery of better advanced and structural materials. These, in turn, can help alleviate needs of several sectors that range from health to energy to the environment. Scientific progress is relatively slow in comparison to the timescale of the problems at hand. It is widely recognized that we need to undertake massive immediate action in the timescale of a decade to enable a transition to a sustainable economy and energy systems to avoid catastrophic problems related to climate change.

The current COVID19 crisis underscores the much-needed agility to solve problems that relate to molecules and materials in the area of health. Our laboratory’s current research aims to accelerate scientific discovery by integrating several disciplines that used to collaborate with each other, but that in our opinion now should merge into a new field of accelerated science. These disciplines include traditional chemistry and materials science, artificial intelligence, data science, analytical and physical chemistry, and robotics. By concentrating on the workflow of scientific discovery and optimization, the concept of a materials acceleration platform or self-driving lab emerges. In this talk, I will describe the components of these platforms and our own efforts to either build them in our laboratory or collaborate with others. I will describe examples in the areas of organic materials and process optimization for the production of pharmaceuticals.

Alán Aspuru-Guzik is a professor of Chemistry and Computer Science at the University of Toronto since July 1st, 2018. He is also the Research Chair of Canada 150 in Theoretical Chemistry and a Canada CIFAR AI Chair at the Vector Institute. Alán began his career at Harvard University in 2006 and was a Full Professor at Harvard University from 2013-2018.

He conducts research in the interfaces of quantum information, chemistry, and machine learning. He was a pioneer in the development of algorithms and experimental implementations of quantum computers and quantum simulators dedicated to chemical systems. He has worked on molecular representations and generative models for the automatic learning of molecular properties. Currently, Alán is interested in automation and "autonomous" chemical laboratories.

He is the recipient of the Google Focused Award for Quantum Computing, the Sloan Research Fellowship, the Camille and Henry Dreyfus Teacher-Scholar award and an Early Career Award in Theoretical Chemistry from the American Chemical Society. In 2010 the MIT Technology Review selected Alán as one of the best innovators under the age of 35. He is an elected member of the American Association for the Advancement of Science (AAAS).

Aspuru-Guzik received his B.Sc. from the National Autonomous University of Mexico (UNAM) in 1999 and obtained a Ph.D. from the University of California, Berkeley in 2004.

None scheduled  

GRADE: Machine-Learning Support for Graduate Admissions 

GRADE is a statistical machine learning system developed to support the work of the graduate admissions committee at the University of Texas at Austin Departmentof Computer Science (UTCS). In recent years, the number of applications to the UTCS PhD program has become too large to manage with a traditional review process. GRADE uses historical admissions data to predict how likely the committee is to admit each new applicant. It reports each prediction as a score similar to those used by human reviewers, and accompanies each by an explanation of what applicant features most influenced its prediction. GRADE makes the review process more efficient by enabling reviewers to spend most of their time on applicants near the decision boundary and by focusing their attention on parts of each applicant’s file that matter the most. An evaluation over two seasons of PhD admissions indicates that the system leads to dramatic time savings, reducing the total time spent on reviews by at least 74%.  Link to the published paper: https://doi.org/10.1609/aimag.v35i1.2504

Zoom: https://umd.zoom.us/j/97904505754?pwd=L2VJYjVVWUN2QVJrOUR0SjVuUFJHdz09

Meeting ID: 979 0450 5754 Passcode: 89593 

TBA