Jimmy Royer

Potential Competition and the Prices of Network Goods: Desktop Software

Potential competition restrains the prices of an incumbent seller when the incumbent can alter the environment perceived by an entrant in a way that both discourages entry and lowers prices. When the product has network effects, the incumbent can make its product ubiquitous and place the potential entrant at a disadvantage because customers have experience with the incumbents product. A primary tool for making a product ubiquitous is low pricing. Hence potential competition lowers the prices of network products. We develop a quantitative model for the desktop software business embodying these principles.

Is backdating executive stock options always costly to shareholders

We use a binomial model to investigate the cost to shareholders of backdating employee stock option (ESO) grants to award in‐the‐money rather than at‐the‐money options to a manager. When the expected return of the stock underlying an ESO is sufficiently close to the risk‐free rate, a backdating arrangement can always be structured to simultaneously improve shareholders’ wealth and the managers utility. The smaller the managers non‐option wealth, personal income tax rate or risk tolerance, the more likely a backdating arrangement can be welfare improving.

Deep Learning Predicts Tuberculosis Drug Resistance Status from Whole-Genome Sequencing Data

Background The diagnosis of multidrug resistant and extensively drug resistant tuberculosis is a global health priority. Whole genome sequencing of clinical Mycobacterium tuberculosis isolates promises to circumvent the long wait times and limited scope of conventional phenotypic antimicrobial susceptibility, but gaps remain for predicting phenotype accurately from genotypic data. Methods and Findings Using targeted or whole genome sequencing and conventional drug resistance phenotyping data from 3,601 Mycobacterium tuberculosis strains, 1,228 of which were multidrug resistant, we investigated the use of machine learning to predict phenotypic drug resistance to 10 anti-tuberculosis drugs. The final model, a multitask wide and deep neural network (MD-WDNN), achieved improved high predictive performance: the average AUCs were 0.979 for first-line drugs and 0.936 for second-line drugs during repeated cross-validation. On an independent validation set, the MD-WDNN showed average AUCs, sensitivities, and specificities, respectively, of 0.937, 87.9%, and 92.7% for first-line drugs and 0.891, 82.0% and 90.1% for second-line drugs. In addition to being able to learn from samples that have only been partially phenotyped, our proposed multidrug architecture shares information across different anti-tuberculosis drugs and genes to provide a more accurate phenotypic prediction. We use t-distributed Stochastic Neighbor Embedding (t-SNE) visualization and feature importance analyses to examine inter-drug similarities. Conclusions Machine learning is capable of accurately predicting resistant status using genomic information and holds promise in bringing sequencing technologies closer to the bedside.

Over-Declaration of Standard Essential Patents and Determinants of Essentiality

Not all Standard Essential Patents (SEPs) are actually essential – a phenomenon called over-declaration. IPR policies of standard-setting organizations require patent holders to declare any patents as SEPs that might be essential, without further SSO review or detailed compulsory declaration information. We analyze actual essentiality of 4G cellular standard SEPs. A declaration against a specific technical specification document of the standard is a strong predictor of essentiality. We also find that citations from and to SEPs declared to the same standard predict essentiality. Our results provide policy guidance and call for recognition of over-declaration in the economics literature.

Decision Making with Machine Learning in Our Modern, Data-Rich Health-Care Industry

Recent innovation in the health-care industry has given us an abundance of data with which we can compare the efficacy of alternative treatments, drugs, and other health interventions. Machine learning has proven to be particularly adept at finding intricate relationships within large datasets. In this chapter we emphasize the potential for machine learning to help us digest and use health-care data effectively. We first provide an introduction to machine learning algorithms, particularly neural network and ensemble algorithms. We then discuss machine learning applications in three areas of the health-care industry. Learning algorithms have been used within the lab as a method of automation to complement problem solving and decision making in the workplace. They have been used to compare the effectiveness of alternative interventions, such as drugs taken together. Given the rise in genomic data, they have been used to develop new treatments and drugs. Taken together, these trends suggest there is vast potential for the expanded application of these algorithms in health care.