MaxMachine interprets product attributes in Amazon's large catalogue.
problem Predicting attribute applicability in large-scale product data with limited ground truth.
method Developed a probabilistic latent variable model that learns distributed binary representations.
result Improves over baseline in 17 out of 19 product groups.
Machine learning analysis of galaxy catalogues finds only one class separable.
problem Difficulty in classifying galaxies using visual inspection schemes.
method Generalized Relevance Matrix Learning Vector Quantization and Random Forests.
result Only one class, Little Blue Spheroids, is consistently separable.
New method extracts cosmological information from dark matter halo catalogues using graph neural networks.
problem Quantifying cosmological information from large-scale structure data.
method Implicit likelihood approach with Information Maximising Neural Networks (IMNNs) on graph representations of dark matter halo catalogues.
result Graph neural network summaries can extract information from noisy catalogues and improve parameter constraints.
Paper speeds up policy optimization for large recommendation systems.
problem Offline optimization of large-scale recommendation systems is computationally expensive.
method Derives an approximation of policy learning algorithms that scales logarithmically with the catalogue size.
result Our algorithm is an order of magnitude faster than naive approaches while producing equally good policies.
The present paper follows the computational approach to 3-manifold classification via edge-coloured graphs, already performed by several authors with respect to orientable 3-manifolds up to 28 coloured tetrahedra, non-orientable 3-manifolds up to 26 coloured tetrahedra, genus two 3-manifolds up to 34 coloured tetrahedr…
Deep learning predicts mismatched ratings in Amazon reviews.
problem Identifying reviews with mismatched ratings on Amazon.
method Converted reviews to vectors using paragraph vector, trained a recurrent neural network with gated recurrent unit, incorporated semantic relationships.
result Model accurately predicts rating mismatches and provides feedback.
Model predicts short-term Amazon rainforest fires with high accuracy.
problem Accurate short-term forecasting of Amazon rainforest fires is challenging.
method Used Seasonal and Trend decomposition based on Loess combined with multi-month-ahead load forecasting algorithms.
result Proposed decomposition-ensemble models provide more accurate forecasts than other models.
The paper models earthquake frequency-magnitude distribution using asymmetric Laplace mixture models.
problem Describing the complete earthquake frequency-magnitude distribution above a completeness magnitude.
method Proposes an asymmetric Laplace mixture model (GFMD-ALMM) to estimate parameters and retrieve mc distribution.
result GFMD-ALMM can accurately model different FMD shapes in various catalogues and sequences.
Amazon SageMaker Model Monitor detects drift in deployed ML models.
problem Ensuring high performance of ML models in production environments.
method Automatically detects data, concept, bias, and feature attribution drift in real-time.
result Maintains high quality models by providing alerts and corrective actions.
Amazon SageMaker AMT automates machine learning model tuning.
problem Challenging hyperparameter tuning for complex machine learning systems.
method Gradient-free optimization using random search or Bayesian optimization.
result AMT finds the best hyperparameter configurations for models.
We describe an algorithm to subdivide automatically a given set of PL n-manifolds (via coloured triangulations or, equivalently, via crystallizations) into classes whose elements are PL-homeomorphic. The algorithm, implemented in the case n=4, succeeds to solve completely the PL-homeomorphism problem among the catalogu…
SFC aims to protect the Amazon with a digital currency and smart contracts.
problem Protecting the Amazon's ecosystem and ensuring resource credibility.
method Blockchain, digital contracts, smart contracts with oracles.
result Ensures credibility and security for financial resources invested in Amazon projects.
The paper uses data science to predict stock trends of Amazon, Apple, Google, and Microsoft.
problem Short-term market movement prediction for major tech stocks.
method Combination of technical analysis and machine/deep learning for trend classification.
result Generated labels for data set: +1 (buy), 0 (hold), -1 (sell).
ASKAP observes a region of the Galactic plane, identifying 3963 radio sources.
problem Characterizing radio sources in the Galactic plane using ASKAP observations.
method 912 MHz observations with 15 ASKAP antennas, source characterization using CAESAR.
result Differential source counts in agreement with previous surveys, spectral index estimation for sources.
Upper bound found for graph manifold complexity.
problem Determining complexity of graph manifolds.
method Using Matveev complexity for all catalogued manifolds.
result Upper bound is sharp for 14,502 catalogued manifolds.
Shared task on machine translation using weak user feedback.
problem Learning machine translation from unreliable user feedback.
method Bandit learning approach using real-valued quality estimates.
result Various machine translation architectures and protocols were evaluated.
Through computer enumeration with the aid of topological results, we catalogue all 18 closed non-orientable P^2-irreducible 3-manifolds that can be formed from at most eight tetrahedra. In addition we give an overview as to how the 100 resulting minimal triangulations are constructed. Observations and conjectures are d…
Drawing together techniques from combinatorics and computer science, we improve the census algorithm for enumerating closed minimal P^2-irreducible 3-manifold triangulations. In particular, new constraints are proven for face pairing graphs, and pruning techniques are improved using a modification of the union-find alg…
We introduce Quintessence: a family of burr puzzles based on the geometry and combinatorics of the 120-cell. We discuss the regular polytopes, their symmetries, the dodecahedron as an important special case, the three-sphere, and the quaternions. We then construct the 120-cell, giving an illustrated survey of its geome…
X-Transformer improves deep transformer performance on extreme multi-label text classification.
problem Classifying text with a large number of labels from a sparse label space.
method Fine-tuning deep transformer models for extreme multi-label text classification.
result X-Transformer achieves new state-of-the-art results on XMC benchmark datasets.
Optimizes binary rating systems for item ranking.
problem Designing efficient feedback systems for item ranking.
method Formalizes performance, provides algorithm, empirically designs and validates.
result Empirically designed and validated approximately optimal rating system.
We have discovered 12 independent new empirical scaling laws in foreign exchange data-series that hold for close to three orders of magnitude and across 13 currency exchange rates. Our statistical analysis crucially depends on an event-based approach that measures the relationship between different types of events. The…
A new framework predicts multi-step time series with stability and efficiency.
problem General probabilistic multi-step time series regression challenges.
method Sequence-to-Sequence Neural Networks, Quantile Regression, Direct Multi-Horizon Forecasting, Forking-sequences training scheme.
result Framework accommodates temporal and static covariates, learning across multiple series.
By means of a slight modification of the notion of GM-complexity, the present paper performs a graph-theoretical approach to the computation of (Matveev's) complexity for closed orientable 3-manifolds. In particular, the existing crystallization catalogue C^{28}, due to Lins, is used to obtain upper bounds for the comp…
MILABOT is a deep reinforcement learning chatbot trained on crowdsourced and user data.
problem Developing a conversational chatbot for small talk topics.
method Deep reinforcement learning applied to an ensemble of natural language models.
result MILABOT outperformed other chatbots in A/B testing with real users.
Study estimates 163 million online freelancers globally.
problem Estimating the number of online workers globally.
method Combining data from various online labour platforms.
result Headline estimate of 163 million registered profiles.
Study improves product categorization on Amazon using multi-modal fusion.
problem Multi-label product categorization in e-commerce.
method Late fusion of image, description, and title modalities using modified CNN and ResNet-50 models.
result Tri-modal late fusion model achieved an F1 score of 88.2%, significantly better than single modal models. As Internet-based commerce becomes increasingly widespread, large data sets about the demand for and pricing of a wide variety of products become available. These present exciting new opportunities for empirical economic and business research, but also raise new statistical issues and challenges. In this article, we su…
Automatic classification of scientific articles based on common characteristics is an interesting problem with many applications in digital library and information retrieval systems. Properly organized articles can be useful for automatic generation of taxonomies in scientific writings, textual summarization, efficient…
Analyzes fashion demand to optimize style mix in retail.
problem Optimizing fashion assortment planning in retail.
method Derives a style quotient from customer demand data, decoupling style from price.
result Shows the relationship between customer perception and retailer's style mix.
GAN-based data augmentation can perpetuate biases in synthetic data.
problem Biases in synthetic data generated by GANs.
method Used a dataset of engineering researchers' head-shots to demonstrate how GANs can reinforce and amplify biases.
result GAN-based data augmentation can amplify biases in synthetic data.
Sparse DNNs face scalability issues; MIT/IEEE/Amazon challenge analyzes best solutions.
problem Scalability issues in Sparse Deep Neural Networks (DNNs).
method Mathematically defined DNN inference computation, community submissions from various fields.
result Sparse DNN execution time, TmDNN, is strongly dependent on the number of operations, Nmop. Satyam simplifies groundtruth collection for machine vision tasks.
problem Difficulty in collecting accurate groundtruth for machine vision systems.
method Crowdsourcing through Amazon Mechanical Turk, automation of task creation and quality control.
result Groundtruth collected by Satyam is comparable to that of trained experts and provides equivalent ML performance.
MACH reduces memory usage for extreme classification by hashing.
problem Expensive training of deep models with large softmax layers.
method Merged-Average Classifiers via Hashing (MACH) using count-min sketch.
result Significant memory reduction and training speedup.
Study compares BERT and XLNet for multi-class categorization of product descriptions.
problem Robustness of multi-class categorization using pre-trained contextualized language models.
method Fine-tuning BERT and XLNet on Amazon product data for multi-class classification.
result Performance decreases linearly with the number of class labels, with BERT consistently outperforming XLNet.
GCNs help in diagnosing label scarcity and feature quality on graphs.
problem Understanding when GCNs improve node classification.
method Simulated label scarcity, feature ablation, and per-class analysis.
result GCNs provide largest gains under extreme label scarcity, matching original performance with noisy features, but hurt when homophily is low and features are strong.
Transformer models outperform recurrent ones in modeling hierarchical data.
problem Modeling hierarchical structure in data.
method Introducing Multiresolution Transformer Networks leveraging self-attention.
result Multiresolution Transformer Networks significantly outperform state-of-the-art models on query suggestion datasets.
Gradient coding improves distributed learning by tolerating failures.
problem Mitigating stragglers in distributed learning.
method Carefully replicating data blocks and coding across gradients.
result Shows tolerance to failures and stragglers for Synchronous Gradient Descent.
We explore how to improve machine translation systems by adding more translation data in situations where we already have substantial resources. The main challenge is how to buck the trend of diminishing returns that is commonly encountered. We present an active learning-style data solicitation algorithm to meet this c…
Developing a visual platform for faster astronomical source cataloging.
problem Speeding up cataloging of large area surveys in radio astronomy.
method Integration of advanced source finding and classification tools into a visual analytic platform.
result Improvement and acceleration of cataloging process in astronomical surveys.
ASOS improves fashion product recommendations by learning from unstructured data.
problem Lack of consistent product information in e-commerce.
method Developed a hybrid recommender system to learn product attributes from unstructured data.
result Quantitative understanding of products improves recommendation accuracy.
Hybrid approach combines user feedback and machine learning for predicting user satisfaction.
problem Measuring user satisfaction in large-scale conversational agent systems.
method Fusion of explicit user feedback and predictions from two machine-learned models trained on different data types.
result Hybrid approach significantly improves user satisfaction predictions.
Annotation errors can significantly hurt classifier performance, yet datasets are only growing noisier with the increased use of Amazon Mechanical Turk and techniques like distant supervision that automatically generate labels. In this paper, we present a robust extension of logistic regression that incorporates the po…
Within crystallization theory, two interesting PL invariants for d-manifolds have been introduced and studied, namely {\it gem-complexity} and {\it regular genus}. In the present paper we prove that, for any closed connected PL 4-manifold M, its gem-complexity k(M) and its regular genus $ \mathcal G(M)…
Efficiently selects data points for deep learning using smaller proxy models.
problem Expensive data selection methods in deep learning.
method Uses a small proxy model to perform data selection.
result Improves data selection runtime by an order of magnitude without significant accuracy loss.
Proposes a method to incorporate prior domain knowledge into hierarchical clustering.
problem Hierarchical clustering results depend on similarity measures and algorithm choices.
method Uses ultrametric distance function to encode external ontological information and adds it as a penalty term to the original pairwise distance.
result Popular linkage-based algorithms can faithfully recover the encoded structure.
This review discusses solutions to Einstein's equations using twistor theory.
problem Finding solutions to Einstein's vacuum equations using twistor theory.
method Holomorphic vector bundles on twistor space and patching matrices.
result Holomorphic patching matrix P is simpler than the metric and determines the rod structure. We present a new method of estimating the distribution of sales rates of, e.g., book titles at an online bookstore, from the time evolution of ranking data found at websites of the store. The method is based on new mathematical results on an infinite particle limit of the stochastic ranking process, and is suitable for…