observations and ideas for improvements  Don’t be afraid of throw away prototypes that don’t work  Good Research Fires Itself  Initial experiments give directions for further improvements  Problem-oriented Spark, TensorFlow, PyTorch  Problem of Weak Baselines  Authors want to show improvements  Successive improvements over state-of-the-art don’t add up Experiments and Result Presentation [Timothy [Timothy G. Armstrong, Alistair Moffat, William Webber, Justin Zobel: Improvements That Don't Add Up: Ad-Hoc Retrieval Results Since 1998. CIKM 2009] [Maurizio Ferrari Dacrema, Paolo Cremonesi, Dietmar Jannach:

experiments during survey / prototype building  Systematic experiments  observations and ideas for improvements  Don’t be afraid of throw away prototypes that don’t work  Ex. Compressed Linear Algebra  % 4. Say what follows from your solution We thereby obtain significant end-to-end performance improvements up to 26x or reduced memory requirements. Structure of Scientific Papers [Simon Peyton Jones:

that we use already, allows shorthands for some operations, and has some performance and safety improvements. 49 The C++11 standard is substantially more complex than its predecessor (1,300 pages versus