“Best before, often good after”

Before After Before After

When Good Enough Is Best

In this issue of Neuron, Rinberg et al. show that mice use a speed-accuracy tradeoff in odor discrimination. Shorter sampling results in high performance for easy problems, and enforced longer sampling results in higher accuracy for difficult problems, but mice freely choose intermediate sampling durations and accuracy varies with difficulty. Reward value and task requirements may determine sam...

Why the Best YA Fiction Often Defies Classification

Afew years ago, I received a phone call from a desperate sixth grade reading teacher. “Help!” she cried, “I have a literary mutiny on my hands. I need your help now!!” I immediately raced upstairs. Our sixth graders read Louis Sachar’s Holes as a required novel. The teacher uses Holes as part of her unit on fantasy. In a time where many middle schoolers are steeped in Harry Potter and Paolini, ...

Voting Early and Often Can Be a Good Thing

The current political climate has almost ruled out the use of Internet voting. Many politicians, led by vocal computer scientists, are pushing for voter verified paper receipts; which is likely to push us even further away from even electronic voting systems. On the other hand, cryptographers have created homomorphic encryption and non-interactive zero-knowledge proofs with features that can su...

Good, Better, Best: Choosing Word Embedding Context

We propose two methods of learning vector representations of words and phrases that each combine sentence context with structural features extracted from dependency trees. Using several variations of neural network classifier, we show that these combined methods lead to improved performance when used as input features for supervised term-matching.