Four short links: 18 August 2016
- Why Should I Trust You? (gitxiv) — In this work, we propose LIME, a novel explanation technique that explains the predictions of any classifier in an interpretable and faithful manner, by learning an interpretable model locally around the prediction.
- A Persona-Based Neural Conversation Model — We present persona-based models for handling the issue of speaker consistency in neural response generation. A speaker model encodes personas in distributed embeddings that capture individual characteristics such as background information and speaking style. A dyadic speaker-addressee model captures properties of interactions between two interlocutors. (via Ben Lorica)
- Randomized Field Experiment in Attracting Early Stage Investors — The average investor responds strongly to information about the founding team, but not to firm traction or existing lead investors. We provide suggestive evidence that team is not merely a signal of quality, and that investing based on team information is a rational strategy. Altogether, our results indicate that information about human assets is causally important for the funding of early-stage firms, and hence, for entrepreneurial success.