📰 Paper reading club: love impossibility results, transformer computational models

We meet biweekly to discuss advances in computer science, ground them down to applications, and see where to the wind is blowing. Read the papers at your own time and come discuss them with us. The talks will be given in English, discussion in En/Ru.

1. 12:00 Love impossibility results (@add512)

It's a simple empirical observation that finding love is hard. We discuss classic and recent works in statistical modelling shedding light on the subject.

- Why I Will Never Have a Girlfriend (Miller, German AI Center, 1999)
- Why I Don't Have a Girlfriend: an Application of the Drake Equation to Love in the UK (Backus, Warwick Economic Summit, 2011)

2. 13:00 RASP/Tracr: a computational model and DSL for writing programs compiled to transformers (@add512)

It's useful to have a computational model for a DL architecture: it allows us to reverse engineer AI and reason about the complexity of algorithms it could learn. RASP is a computational model for transformers in the same way automata are a computational model for RNNs; Tracr is a DSL you can write programs that compile into transformers in.

This is exciting in terms of PL/DL connections and suggests an exciting interpretability agenda.

⏱ 12:00-14:30 Saturday, 30 Mar
📍 F0RTHSP4CE, Khorava St, 18
🗓 no registration/just come
💰 free/donation/boost
👅 en+ru
👮 @add512
💬 event chat and q&a
🗣 propose talk
🎥 recordings