Version Of Machine Learning Research Articles

Cells solved the Gibbs paradox by learning to contain entropic forces

As Nature’s version of machine learning, evolution has solved many extraordinarily complex problems, none perhaps more remarkable than learning to harness an increase in chemical entropy (disorder) to generate directed chemical forces (order). Using muscle as a model system, here I describe the basic mechanism by which life creates order from disorder. In short, evolution tuned the physical properties of certain proteins to contain changes in chemical entropy. As it happens these are the “sensible” properties Gibbs postulated were needed to solve a paradox that has intrigued and challenged scientists and philosophers for over 100 years.

Machine Learning Optimization of Quantum Circuit Layouts

The quantum circuit layout (QCL) problem involves mapping out a quantum circuit such that the constraints of the device are satisfied. We introduce a quantum circuit mapping heuristic, QXX, and its machine learning version, QXX-MLP. The latter automatically infers the optimal QXX parameter values such that the laid out circuit has a reduced depth. In order to speed up circuit compilation, before laying the circuits out, we use a Gaussian function to estimate the depth of the compiled circuits. This Gaussian also informs the compiler about the circuit region that influences most the resulting circuit’s depth. We present empiric evidence for the feasibility of learning the layout method using approximation. QXX and QXX-MLP open the path to feasible large-scale QCL methods.

Analysis of the Effectiveness of Machine Learning in Determining Decision Rules for Executive Compensation Planning

This paper describes the use of a machine-learning technique to derive executive compensation planning rules. The specific learning technique was AQ15, which constructs rules by classifying sample data according to features or characteristics. The validity of the resulting rules was tested by obtaining a measure of performance called an estimate of the error rate. The machine-derived rules were then compared to rules made by human experts and by two other versions of AQ15. Although not statistically significant, the machine-learning version of AQ15 using hypotheses from human experts outperformed the other methods of rule construction.