Meta-learning neural Bloom filters

Meta-learning neural bloom filters Rae et al., ICML'19 Bloom filters are wonderful things, enabling us to quickly ask whether a given set could possibly contain a certain value. They produce this answer while using minimal space and offering O(1) inserts and lookups. It’s no wonder Bloom filters and their derivatives (the family of approximate set … Continue reading Meta-learning neural Bloom filters

Challenging common assumptions in the unsupervised learning of disentangled representations

Challenging common assumptions in the unsupervised learning of disentangled representations Locatello et al., ICML'19 Today’s paper choice won a best paper award at ICML’19. The ‘common assumptions’ that the paper challenges seem to be: "unsupervised learning of disentangled representations is possible, and useful!" The key idea behind the unsupervised learning of disentangled representations is that … Continue reading Challenging common assumptions in the unsupervised learning of disentangled representations

Data Shapley: equitable valuation of data for machine learning

Data Shapley: equitable valuation of data for machine learning Ghorbani & Zou et al., ICML'19 It’s incredibly difficult from afar to make sense of the almost 800 papers published at ICML this year! In practical terms I was reduced to looking at papers highlighted by others (e.g. via best paper awards), and scanning the list … Continue reading Data Shapley: equitable valuation of data for machine learning

Software engineering for machine learning: a case study

Software engineering for machine learning: a case study Amershi et al., ICSE'19 Previously on The Morning Paper we’ve looked at the spread of machine learning through Facebook and Google and some of the lessons learned together with processes and tools to address the challenges arising. Today it’s the turn of Microsoft. More specifically, we’ll be … Continue reading Software engineering for machine learning: a case study

Machine learning systems are stuck in a rut

Machine learning systems are stuck in a rut Barham & Isard, HotOS'19 In this paper we argue that systems for numerical computing are stuck in a local basin of performance and programmability. Systems researchers are doing an excellent job improving the performance of 5-year old benchmarks, but gradually making it harder to explore innovative machine … Continue reading Machine learning systems are stuck in a rut

A case for managed and model-less inference serving

A case for managed and model-less inference serving Yadwadkar et al., HotOS'19 HotOS’19 is presenting me with something of a problem as there are so many interesting looking papers in the proceedings this year it’s going to be hard to cover them all! As a transition from the SysML papers we’ve been looking at recently, … Continue reading A case for managed and model-less inference serving

PyTorch-BigGraph: a large-scale graph embedding system

PyTorch-BigGraph: a large-scale graph embedding system Lerer et al., SysML'19 We looked at graph neural networks earlier this year, which operate directly over a graph structure. Via graph autoencoders or other means, another approach is to learn embeddings for the nodes in the graph, and then use these embeddings as inputs into a (regular) neural … Continue reading PyTorch-BigGraph: a large-scale graph embedding system