Therefore, due to a higher morbidity in REM dependent OSA it may be justified to consider CPAP treatment at lower AHI levels.

"the wealth tax is likely to be the most direct and powerful tool to restore tax progressivity at the very top of the distribution"

"our analysis shows that the wealth tax has great revenue and wealth equalizing
potential in the US context. The wealth tax, if the tax rates are high enough, is also a powerful tool to deconcentrate wealth. A wealth tax of 2 or 3% per year can put a significant dent into this growth rate advantage. With successful enforcement, a wealth tax has to deliver either revenue or de-concentrate wealth.79 Set the rates low (1%) and you get revenue in perpetuity but little (or very slow) de-concentration. Set the rates medium (2-3%) and you get revenue for quite a while and de-concentration eventually. Set the rates high (significantly above 3%) and you get de-concentration fast but revenue does not last long. Which is best will depend of course on ones political views."

"Can a wealth tax be successfully enforced? Our review of past and foreign experiences, as well as recent empirical work tells us that enforcement is a policy choice."

"The wealth tax accelerates the process of dispersion of stock ownership for very successful businesses that make their owners-founders billionaires. Dispersed stock ownership has been a feature of US capitalism and is a key reason why taxing wealthy business owners is feasible. Importantly and in contrast to labor income, this dispersion does not mean that economic activity disappears. There might not be even any effect on the wealth stock if the government uses the wealth tax proceeds for public investment, debt reduction, or to create a sovereign fund. The wealth disappears only if the government cannot save the money and cannot encourage middle class saving."

there are four principal exposures that explain up to 76% of corporate-debt returns, Israelov calculates: government obligations, equities, stock volatility and price swings in bonds. In his parlance, these are the most-rewarded risks out there for credit buyers.

In that spirit, investors can garner exposure to the asset class via a portfolio of fixed-income and equity-index futures, combined with selling options on a stock index and bond futures, according to the paper. All without holding cash bonds -- with smaller drawdowns and lower volatility compared with benchmarks.

paper: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3293357

The back-propagation algorithm is the cornerstone of deep learning. Despite its importance, few variations of the algorithm have been attempted. This work presents an approach to discover new variations of the back-propagation equation. We use a domain specific lan- guage to describe update equations as a list of primitive functions. An evolution-based method is used to discover new propagation rules that maximize the generalization per- formance after a few epochs of training. We find several update equations that can train faster with short training times than standard back-propagation, and perform similar as standard back-propagation at convergence.

accepted papers at:
BayesOpt 2017
NIPS Workshop on Bayesian Optimization
December 9, 2017
Long Beach, USA

We introduced ROBO, a flexible Bayesian optimization framework in python. For standard GP-based
blackbox optimization, its performance is on par with Spearmint while using the permissive BSD
license. Most importantly, to the best of our knowledge, ROBO is the first BO package that includes
Bayesian neural network models and that implements specialized BO methods that go beyond the
blackbox paradigm to allow orders of magnitude speedup.

Because foreign-law bonds are often priced in a foreign currency, we need to adjust the observed yields for the currency premium