Heuristics ≈ Heuristics
View article
SMS: Smart Model Selection in PhyML Open
Model selection using likelihood-based criteria (e.g., AIC) is one of the first steps in phylogenetic analysis. One must select both a substitution matrix and a model for rates across sites. A simple method is to test all combinations and …
View article
Sample Size Justification Open
An important step when designing an empirical study is to justify the sample size that will be collected. The key aim of a sample size justification for such studies is to explain how the collected data is expected to provide valuable info…
View article
Machine learning for combinatorial optimization: A methodological tour d'horizon Open
This paper surveys the recent attempts, both from the machine learning and operations research communities, at leveraging machine learning to solve combinatorial optimization problems. Given the hard nature of these problems, state-of-the-…
View article
Learning Combinatorial Optimization Algorithms over Graphs Open
The design of good heuristics or approximation algorithms for NP-hard combinatorial optimization problems often requires significant specialized knowledge and trial-and-error. Can we automate this challenging, tedious process, and learn th…
View article
QSGD: Communication-Efficient SGD via Gradient Quantization and Encoding Open
Parallel implementations of stochastic gradient descent (SGD) have received significant research attention, thanks to excellent scalability properties of this algorithm, and to its efficiency in the context of training deep neural networks…
View article
Right for the Wrong Reasons: Diagnosing Syntactic Heuristics in Natural Language Inference Open
A machine learning system can score well on a given test set by relying on heuristics that are effective for frequent example types but break down in more challenging cases. We study this issue within natural language inference (NLI), the …
View article
Snorkel Open
Labeling training data is increasingly the largest bottleneck in deploying machine learning systems. We present Snorkel, a first-of-its-kind system that enables users to train state-of-the-art models without hand labeling any training data…
View article
From the Quantum Approximate Optimization Algorithm to a Quantum Alternating Operator Ansatz Open
The next few years will be exciting as prototype universal quantum processors emerge, enabling the implementation of a wider variety of algorithms. Of particular interest are quantum heuristics, which require experimentation on quantum har…
View article
This Just In: Fake News Packs A Lot In Title, Uses Simpler, Repetitive Content in Text Body, More Similar To Satire Than Real News Open
The problem of fake news has gained a lot of attention as it is claimed to have had a significant impact on 2016 US Presidential Elections. Fake news is not a new problem and its spread in social networks is well-studied. Often an underlyi…
View article
Interacting with educational chatbots: A systematic review Open
Chatbots hold the promise of revolutionizing education by engaging learners, personalizing learning activities, supporting educators, and developing deep insight into learners’ behavior. However, there is a lack of studies that analyze the…
View article
Learning scheduling algorithms for data processing clusters Open
Efficiently scheduling data processing jobs on distributed compute clusters requires complex algorithms. Current systems use simple, generalized heuristics and ignore workload characteristics, since developing and tuning a scheduling polic…
View article
Automated optimized parameters for T-distributed stochastic neighbor embedding improve visualization and analysis of large datasets Open
Accurate and comprehensive extraction of information from high-dimensional single cell datasets necessitates faithful visualizations to assess biological populations. A state-of-the-art algorithm for non-linear dimension reduction, t-SNE, …
View article
To Trust or to Think Open
People supported by AI-powered decision support tools frequently overrely on the AI: they accept an AI's suggestion even when that suggestion is wrong. Adding explanations to the AI decisions does not appear to reduce the overreliance and …
View article
'It's Reducing a Human Being to a Percentage' Open
Data-driven decision-making consequential to individuals raises important questions of accountability and justice. Indeed, European law provides individuals limited rights to 'meaningful information about the logic' behind significant, aut…
View article
Artificial Intelligence and Surgical Decision-making Open
Integration of artificial intelligence with surgical decision-making has the potential to transform care by augmenting the decision to operate, informed consent process, identification and mitigation of modifiable risk factors, decisions r…
View article
Massive Exploration of Neural Machine Translation Architectures Open
Neural Machine Translation (NMT) has shown remarkable progress over the past few years, with production systems now being deployed to end-users. As the field is moving rapidly, it has become unclear which elements of NMT architectures have…
View article
Developing prediction models for clinical use using logistic regression: an overview Open
Prediction models help healthcare professionals and patients make clinical decisions. The goal of an accurate prediction model is to provide patient risk stratification to support tailored clinical decision-making with the hope of improvin…
View article
Real options theory in strategic management Open
Research summary : T his article provides a review of real options theory ( ROT ) in strategic management research. We review the fundamentals of ROT and provide a taxonomy of this research. By synthesizing and critiquing research on real …
View article
A Hierarchical Taxonomy of Psychopathology Can Transform Mental Health Research Open
For more than a century, research on psychopathology has focused on categorical diagnoses. Although this work has produced major discoveries, growing evidence points to the superiority of a dimensional approach to the science of mental ill…
View article
Noisy Networks for Exploration Open
We introduce NoisyNet, a deep reinforcement learning agent with parametric noise added to its weights, and show that the induced stochasticity of the agent's policy can be used to aid efficient exploration. The parameters of the noise are …
View article
Graph-to-Sequence Learning using Gated Graph Neural Networks Open
Many NLP applications can be framed as a graph-to-sequence learning problem. Previous work proposing neural architectures on graph-to-sequence obtained promising results compared to grammar-based approaches but still rely on linearisation …
View article
Memory, Attention, and Choice* Open
Building on a textbook description of associative memory (Kahana 2012), we present a model of choice in which a choice option cues recall of similar past experiences. Memory shapes valuation and decisions in two ways. First, recalled exper…
View article
Searching for “Defect-Tolerant” Photovoltaic Materials: Combined Theoretical and Experimental Screening Open
Recently, we and others have proposed screening criteria for 'defect-tolerant' photovoltaic (PV) absorbers, identifying several classes of semiconducting compounds with electronic structures similar to those of hybrid lead-halide perovskit…
View article
AMC: AutoML for Model Compression and Acceleration on Mobile Devices Open
Model compression is a critical technique to efficiently deploy neural network models on mobile devices which have limited computation resources and tight power budgets. Conventional model compression techniques rely on hand-crafted heuris…
View article
Data Programming: Creating Large Training Sets, Quickly Open
Large labeled training sets are the critical building blocks of supervised learning methods and are key enablers of deep learning techniques. For some applications, creating labeled training sets is the most time-consuming and expensive pa…
View article
Training Region-Based Object Detectors with Online Hard Example Mining Open
The field of object detection has made significant advances riding on the wave of region-based ConvNets, but their training procedure still includes many heuristics and hyperparameters that are costly to tune. We present a simple yet surpr…
View article
#Exploration: A Study of Count-Based Exploration for Deep Reinforcement Learning Open
Count-based exploration algorithms are known to perform near-optimally when used in conjunction with tabular reinforcement learning (RL) methods for solving small discrete Markov decision processes (MDPs). It is generally thought that coun…
View article
Distributed deep learning networks among institutions for medical imaging Open
Objective Deep learning has become a promising approach for automated support for clinical diagnosis. When medical data samples are limited, collaboration among multiple institutions is necessary to achieve high algorithm performance. Howe…
View article
Non-convex Optimization for Machine Learning Open
A vast majority of machine learning algorithms train their models and perform\ninference by solving optimization problems. In order to capture the learning\nand prediction problems accurately, structural constraints such as sparsity or\nlo…
View article
Dealing with multiple decades of hourly wind and PV time series in energy models: A comparison of methods to reduce time resolution and the planning implications of inter-annual variability Open
Using a high-resolution planning model of the Great Britain power system and 25 years of simulated wind and PV generation data, this study compares different methods to reduce time resolution of energy models to increase their computationa…