Study examines complexity of opinions in online discussions.
problem Understanding the complexity and similarity of opinions in online discussions.
method Developed a modeling framework to construct latent representations of opinions, connecting them to voting models and sign-rank of a matrix.
result Found that unidimensional models often fail to accurately represent online discussions, revealing insights into human judgements and opinions.
New method predicts political ideology from online activity.
problem Predicting political ideology from digital footprints.
method Statistical learning approaches applied to reddit data.
result Activity in non-political forums can predict political ideology with high accuracy.
In online discussion communities, users can interact and share information and opinions on a wide variety of topics. However, some users may create multiple identities, or sockpuppets, and engage in undesired behavior by deceiving others or manipulating discussions. In this work, we study sockpuppetry across nine discu…
Paper improves probabilistic forecasts of electricity prices.
problem Combining multivariate probabilistic forecasts with dependencies.
method Online learning algorithm using Bernstein Online Aggregation (BOA).
result Significant improvement in CRPS over uniform combination.
We discuss multi-task online learning when a decision maker has to deal simultaneously with M tasks. The tasks are related, which is modeled by imposing that the M-tuple of actions taken by the decision maker needs to satisfy certain constraints. We give natural examples of such restrictions and then discuss a general …
New definition of regret for nonconvex online learning models.
problem Intractability of standard regret measures for nonconvex models.
method Introduced a local gradient based regret definition.
result Our definition provides more interpretable bounds for forecasting.
Online portfolio selection is a fundamental problem in computational finance, which has been extensively studied across several research communities, including finance, statistics, artificial intelligence, machine learning, and data mining, etc. This article aims to provide a comprehensive survey and a structural under…
We consider the problem of learning from noisy data in practical settings where the size of data is too large to store on a single machine. More challenging, the data coming from the wild may contain malicious outliers. To address the scalability and robustness issues, we present an online robust learning (ORL) approac…
In this paper, we introduce a methodology that allows to model behavioral trajectories of users in online social media. First, we illustrate how to leverage the probabilistic framework provided by Hidden Markov Models (HMMs) to represent users by embedding the temporal sequences of actions they performed online. We the…
Massive Open Online Courses (MOOCs) bring together thousands of people from different geographies and demographic backgrounds -- but to date, little is known about how they learn or communicate. We introduce a new content-analysed MOOC dataset and use Bayesian Non-negative Matrix Factorization (BNMF) to extract communi…
Private classification and online prediction are shown to be equivalent.
problem Learning with differential privacy and online prediction equivalence.
method Introducing global stability and proving equivalence between online learnability and private PAC learnability.
result Every concept class with finite Littlestone dimension can be learned by a differentially-private algorithm.
Adversarial online nonparametric regression achieves optimal rates with locally adaptive learning.
problem Adversarial online nonparametric regression with general convex losses.
method Parameter-free learning algorithm leveraging chaining trees to compete against H{ö}lder functions, dynamically tracking and adapting to local smoothness variations.
result First computationally efficient algorithm with locally adaptive optimal rates for online regression in an adversarial setting.
New method improves forecast accuracy using CRPS for probabilistic predictions.
problem Improving forecast accuracy for probabilistic predictions, especially in the tails of distributions.
method Introduces a new weighting method for pointwise CRPS learning, optimizing across quantiles.
result Proposed fully adaptive Bernstein online aggregation (BOA) method for pointwise CRPS online learning has optimal convergence properties.
Adversaries can poison online learning data to mislead models.
problem Adversarial attacks on online learning systems.
method Formalized problem into two settings, proposed optimization-based attack strategy, and three solution strategies.
result Demonstrated effectiveness of data poisoning attacks in online learning.
In the current context of data explosion, online techniques that do not require storing all data in memory are indispensable to routinely perform tasks like principal component analysis (PCA). Recursive algorithms that update the PCA with each new observation have been studied in various fields of research and found wi…
User contributions in the form of posts, comments, and votes are essential to the success of online communities. However, allowing user participation also invites undesirable behavior such as trolling. In this paper, we characterize antisocial behavior in three large online discussion communities by analyzing users who…
Paper proposes an online covariance estimator for sketched Newton methods.
problem Estimating the limiting covariance matrix of sketched Newton methods.
method Proposes a fully online covariance matrix estimator from Newton iterates.
result Establishes the consistency and convergence rate of the proposed estimator.
Paper presents derivative-free methods for online inverse dynamics modeling.
problem Online learning of inverse dynamics models without numerical differentiation.
method Derivative-free framework for rigid body dynamics, data-driven, and semiparametric models.
result Proposed `derivative-free' methods outperform existing methodologies in real data experiments.
We present and analyse three online algorithms for learning in discrete Hidden Markov Models (HMMs) and compare them with the Baldi-Chauvin Algorithm. Using the Kullback-Leibler divergence as a measure of generalisation error we draw learning curves in simplified situations. The performance for learning drifting concep…
Book introduces online learning via convex optimization, focusing on regret minimization.
problem Minimizing regret in online learning under worst-case assumptions.
method Unified view of online learning through convex optimization, including adaptive and parameter-free algorithms.
result Unified understanding of various online learning algorithms and their applications.
Proposes standards for evaluating online machine learning methods in evolving data streams.
problem Difficulty in evaluating online machine learning methods under realistic conditions.
method Proposes comprehensive evaluation standards, performance measures, and evaluation strategies.
result Provides a new Python framework (float) for modular integration of libraries and custom code.
Probabilistic label trees improve XMLC by organizing labels hierarchically.
problem Efficiently tagging instances with a small subset of relevant labels from a large pool.
method Introduce and analyze probabilistic label trees (PLTs) as a generalization of hierarchical softmax for multi-label problems.
result PLTs are consistent for various performance metrics and can be trained online without prior knowledge.
Paper proposes using word embeddings to detect trolls in social media debates.
problem Preventing online harassment through rapid detection of offensive posts.
method Word embedding models for identifying fast-changing topics and negative content.
result GloVe model helps in discovering new keywords for trolling detection.
Neuromorphic column performs online unsupervised clustering.
problem Real-time clustering of streaming data.
method Localized, spike timing-dependent plasticity (STDP) neural column.
result Prototype column performs similarly to k-means clustering.
New algorithms adapt to both gradient norms and comparator norms in online learning.
problem Adapting to both gradient norms and comparator norms in online learning.
method Developed parameter-free and scale-free algorithms for unbounded online convex optimization.
result Improved regret bounds for scale-invariant online prediction with linear models.
New method uses offline data to improve online bandit learning, even when distributions differ.
problem Improving online bandit learning with different offline and online distributions.
method MIN-UCB policy that adapts to offline data when informative, achieving tight regret bounds.
result MIN-UCB policy outperforms UCB policy with offline data and provides tight regret bounds.
Paper proposes an efficient PCA method for memory-limited streaming data.
problem Memory-limited online PCA methods struggle with convergence and eigenvector extraction.
method A single-pass algorithm converging to the first k eigenvectors without predefined parameters.
result Empirical results show the scheme converges to first k eigenvectors in a single pass.
The study analyzes online predictions for non-stationary time series under model misspecification.
problem Analyzing predictive properties of statistical methods in non-stationary time series under model misspecification.
method Defining Kullback-Leibler risk, proving minimax predictive densities for dynamic models, extending results to multiple predictive densities.
result Dynamic random walk models produce exact minimax predictive densities under Gaussian assumptions and semi-martingale processes.
We provide an online RLHF workflow for large language models.
problem Lack of online RLHF workflow for large language models.
method Construct preference models using open-source datasets, approximate human feedback, and implement online iterative RLHF.
result Our LLM achieves state-of-the-art performance on various benchmarks.
This article surveys online machine learning in big data streams.
problem Limited storage for past data in data streams.
method Distributed software architectures and libraries for efficient algorithms.
result Overview of classification, regression, recommendation, and unsupervised models for streaming data.
New framework for tracking varying bounds in time series forecasting.
problem Forecasting bounded time series with varying bounds.
method Extended log-likelihood estimation, online maximum likelihood estimation, Normalized Gradient Descent (NGD) for quasiconvex optimization.
result Derive an Online Normalized Gradient Descent algorithm for online bound tracking.
Trajectory segmentation is the process of subdividing a trajectory into parts either by grouping points similar with respect to some measure of interest, or by minimizing a global objective function. Here we present a novel online algorithm for segmentation and summary, based on point density along the trajectory, and …
Novel online algorithm for hierarchical imitation learning.
problem Scalability issue in reinforcement learning and options discovery.
method Online Baum-Welch algorithm for hierarchical imitation learning.
result The online algorithm outperforms the batch version in both discrete and continuous environments.
Unified approach to fair online learning with stochastic contexts.
problem Fairness in online learning with unknown sensitive contexts.
method Adapting Blackwell's approachability theory to handle unknown contexts' distributions.
result Characterization of optimal trade-off between fairness and performance objectives.
This paper explores online learning of dynamics and state using ensemble Kalman filters.
problem Reconstructing dynamics from partial and noisy observations in real-time.
method Ensemble Kalman filter (EnKF) family of algorithms for online learning of dynamics and state.
result Demonstrates the efficiency and accuracy of online learning methods using Lorenz models.
New methods improve Bayesian inference and decision-making in online learning.
problem Current Bayesian deep learning does not fully utilize joint predictives for sequential decision-making.
method Proposes new evaluation settings for active learning and active sampling, focusing on marginal and joint cross-entropies.
result Initial experiments suggest challenges in applying current BDL inference techniques in high-dimensional spaces.
The study predicts how discussions in mental disorder Reddit communities affect users' emotional states.
problem Improving mental health conditions through social support analysis.
method Text embedding techniques and RNNs for predicting emotional tone shifts.
result Users' emotional states can improve due to social support, as evidenced by positive comments following negative posts.
The paper extends a prediction method to curved spaces.
problem Prediction with expert advice in curved spaces.
method Geometric considerations and adaptation of online to batch conversion.
result Extension of exponentially weighted average forecaster to non-positive curvature spaces.
Group discussions are essential for organizing every aspect of modern life, from faculty meetings to senate debates, from grant review panels to papal conclaves. While costly in terms of time and organization effort, group discussions are commonly seen as a way of reaching better decisions compared to solutions that do…
Optimal sampling reduces power grid data analysis costs.
problem Efficient online analysis of high-speed, correlated IoT data.
method D-optimality criterion-based sampling methods combining Bernoulli and leverage score sampling.
result Leverage score sampling improves computational efficiency and outperforms benchmarks.
The paper proposes a method to predict audio ad quality using acoustic features.
problem Improving user experience in online music streaming services by ensuring high quality audio advertisements.
method The paper proposes predicting audio ad quality using acoustic features and a proxy metric called Long Click Rate (LCR). A deep learning model is also introduced.
result The proposed deep learning model outperforms other models trained on hand-crafted features for audio ad quality prediction.
A method for online tensor dictionary learning is proposed. With the assumption of separable dictionaries, tensor contraction is used to diminish a N-way model of O(LN) into a simple matrix equation of O(NL2) with a real-time capability. To avoid numerical instability d…
Proposes Hebbian-descent for neural network learning, addressing Hebbian and gradient descent issues.
problem Learning issues with correlated data and vanishing error term in gradient descent.
method Integrates Hebbian and gradient descent principles without activation function derivatives, centering neural activities.
result Biologically plausible, convergent, and effective in online learning with correlated data.
Machine learning experiments often mislead due to unmet assumptions.
problem Machine learning experiments with pooled data may not meet necessary assumptions for unbiased causal effect estimation.
method Analysis of assumptions required for unbiased causal effect estimation in machine learning experiments.
result Practical applications of A/B-tests with machine learning models may not yield unbiased estimates of causal effect.
Study detects concept shift in online data using martingales.
problem Detecting concept shift in online datasets.
method Exchangeable martingales and conformal prediction techniques.
result Decomposes concept shift into detectable components.
Blog post discusses various implementations of Fisher Information for EWC in continual learning.
problem Improving Elastic Weight Consolidation (EWC) results by optimizing Fisher Information computation.
method Empirically compares different implementations of Fisher Information for EWC.
result Many reported EWC results can be improved by changing Fisher Information computation methods.
We address online linear optimization problems when the possible actions of the decision maker are represented by binary vectors. The regret of the decision maker is the difference between her realized loss and the best loss she would have achieved by picking, in hindsight, the best possible action. Our goal is to unde…
Paper tackles online PCA, showing non-convex gradient ascent can minimize regret.
problem Online Principal Component Analysis in adversarial settings.
method Adversarially-perturbed spiked-covariance model, non-convex online gradient ascent with warm-start.
result Non-convex online gradient ascent can minimize regret with high probability.