"Chinese Word Segmentation"
Albert Ji, '21
We investigated using machine learning for the word segmentation problem. We trained a model just with Chinese, and then another with several languages other than Chines. We found that a multi-lingual model works much better than a Chinese-only model, that is, training the model with multiple languages improves the performance. Our work suggests that the concept of being a word is somehow transportable across languages.
"Differentially Private Hypothesis Testing"
Kaiyan Shi, '20
A general differentially private framework on hypothesis testing. After subsampling data and performing public hypothesis test on each sub-dataset, we aggregate them differentially privately to obtain a final resulting p-value for the dataset. We adopt this framework on several models to see its power and its quality in application.
"Gradual Verification with Recursive Predicates"
Henry Blanchette, '20
Two common strategies are used to verify that programs meet their specification: proofs at compile time ("static" verification) and assertions at run time ("dynamic" verification). I'll present work on an implementation of gradual verification that allows less precise specification, and thus is a mix of the two. The goal is to offer partial static verification when full static verification is infeasible.
"Differentially Private Confidence Intervals"
Monica Moniot, '20
Tuesday, October 8, 2019 at 4:30pm to 6:00pm
Reed Community Members
If you are a member of the Reed community, you MUST LOG IN to see events that are open ONLY to the Reed community. Log in with your Reed ID (your Kerberos account information). If you don’t remember your account username or password, go to reed.edu/cis/help/kerberos.html.Log in with Reed ID