Transformer from scratch
https://e2eml.school/transformers.html Transformers were introduced in this 2017 paper as a tool for sequence transduction—converting one sequence of symbols to another. The most popular examples of this are translation, as in English to German. It has also been modified to perform sequence completion—given a starting prompt, carry on in the same...
Interleaving: A Classroom Experiment
https://www.learningscientists.org/blog/2021/10/28-1 The researchers (1) were interested in whether interleaving math problems over time would lead to superior learning than blocking them. Importantly, in this experiment, there were all different types of math problems. During this time, the students received their normal lessons and assignments. There were four different types of...
Unpopular opinion - Data scientists should be more end-to-end
https://eugeneyan.com/writing/end-to-end-data-science/ More context, faster iteration, greater satisfaction For most data science roles, being more end-to-end improves your ability to make meaningful impact. (Nonetheless, there are roles that focus on machine learning.) Working end-to-end provides increased context. While specialized roles can increase efficiency, it reduces context (for the data scientist) and...
The sobering truth about the impact of your business ideas
https://www.oreilly.com/radar/the-sobering-truth-about-the-impact-of-your-business-ideas/ But, science-based organizations are rigorously quantifying this impact and have learned some sobering lessons: The vast majority of business ideas fail to generate a positive impact. Most companies are unaware of this. It is unlikely that companies will increase the success rate for their business ideas. For the very...
The sobering truth about the impact of your business ideas
https://openai.com/blog/grade-school-math/ https://arxiv.org/pdf/2110.14168.pdf https://github.com/openai/grade-school-math We’ve trained a system that solves grade school math problems with nearly twice the accuracy of a fine-tuned GPT-3 model. It solves about 90% as many problems as real kids: a small sample of 9-12 year olds scored 60% on a test from our dataset, while our...
People are more tolerant of inequality when it is expressed in terms of individuals rather than groups at the top
https://www.pnas.org/content/118/43/e2100430118 Despite the ever-growing economic gap between the very wealthy and the rest of the population, support for redistributive policies tends to be low. This research tested whether people’s tolerance of inequality differs when it is represented in terms of a successful individual versus a group of people at the...
People mistake the internet’s knowledge for their own
https://www.pnas.org/content/118/43/e2105061118 People frequently search the internet for information. Eight experiments (n = 1,917) provide evidence that when people “Google” for online information, they fail to accurately distinguish between knowledge stored internally—in their own memories—and knowledge stored externally—on the internet. Relative to those using only their own knowledge, people who use...
Development of children’s math attitudes: Gender differences, key socializers, and intervention approaches
https://www.sciencedirect.com/science/article/abs/pii/S0273229721000526 The relation of various math attitudes to math achievement has been extensively studied in adolescents and adults. Recently, researchers have begun to examine the math attitude-math achievement relation in young children. We review theories and research on four attitudes relevant to early math learning—math anxiety, math self-concept, mindset, and...
Research: Cameras On or Off?
https://hbr.org/2021/10/research-cameras-on-or-off Managers looking to encourage engagement and inclusion in remote meetings have long encouraged team members to keep their cameras turned on. But researchers examining remote employees’ reactions to the constant video conference calls of the remote work era have found that keeping video on all day actually increases so-called...
Probability models for customer-base analysis
https://brucehardie.com/talks.html https://brucehardie.com/talks/cba_tut_art_16_HO.pdf Fader, Peter S. and Bruce G.S. Hardie (2007), “How to Project Customer Retention,” Journal of Interactive Marketing, 21 (Winter), 76–90. Fader, Peter S. and Bruce G.S. Hardie (2014), “A Spreadsheet-Literate Non-Statistician’s Guide to the Beta-Geometric Model.” http://brucehardie.com/notes/032/ Fader, Peter S. and Bruce G. S. Hardie (2010), “Customer-Base Valuation...
About
a collection of excerpts from interesting research, blogs, events, etc...