A new update rule for deep reinforcement learning reduces learning variance and variance in reference signals.
problem Learning variance and incorrect reference signals in deep reinforcement learning.
method t-soft update method inspired by student-t distribution, which reduces extreme updates and accelerates similar updates.
result The t-soft update method outperforms conventional methods in terms of return and variance in PyBullet robotics simulations.
Federated learning can be vulnerable to adversarial attacks, which this work addresses.
problem Federated learning's adversarial vulnerability when deployed.
method Bias-Variance decomposition for federated learning, proposing Fed_BVA framework.
result Fed_BVA framework generates adversarial examples to improve robustness.
Improved HGF networks avoid negative precision errors in volatility updates.
problem Negative posterior precision errors in volatility-coupled nodes of HGF networks.
method Introduced a modified quadratic approximation to variational energy.
result Robust update equations across parameter space that track posterior faithfully.
Importance sampling (IS) is a common reweighting strategy for off-policy prediction in reinforcement learning. While it is consistent and unbiased, it can result in high variance updates to the weights for the value function. In this work, we explore a resampling strategy as an alternative to reweighting. We propose Im…
Recently, a lot of effort has been paid to the efficient computation of Kriging predictors when observations are assimilated sequentially. In particular, Kriging update formulae enabling significant computational savings were derived in Barnes and Watson (1992), Gao et al. (1996), and Emery (2009). Taking advantage of …
Humans are able to accelerate their learning by selecting training materials that are the most informative and at the appropriate level of difficulty. We propose a framework for distributing deep learning in which one set of workers search for the most informative examples in parallel while a single worker updates the …
The paper bounds generalization error for iterative learning with bounded updates.
problem Generalization error of iterative learning algorithms with bounded updates for non-convex loss functions.
method Information-theoretic techniques, reformulating mutual information as update uncertainty, variance decomposition.
result Improved generalization error bounds for iterative learning algorithms with bounded updates.
A new algorithm SRG-DQN reduces variance in deep Q-learning.
problem Inaccurate estimation of anchor points in SVRG for deep Q-learning.
method Introduces recursive gradient variance reduction for stochastic gradient updates.
result Demonstrates improved efficiency and effectiveness of SRG-DQN on reinforcement learning tasks.
Paper introduces a new policy optimization method using importance sampling.
problem Stable and low variance policy learning with small policy updates.
method Derives an alternative objective using importance sampling and introduces an approximation to balance bias and variance.
result The new algorithm improves on-policy policy optimization on continuous control benchmarks.
FedGLOMO accelerates FL convergence for non-convex functions.
problem Efficiently solving non-convex optimization problems in federated learning with client heterogeneity.
method Combines global and local momentum updates to reduce variance and improve convergence rate.
result Achieves O(ε−1.5) convergence to ε-stationary point, compared to O(ε−2). Improves inference-time alignment for diffusion models without updating weights.
problem Aligning diffusion models without updating weights for high-reward outputs.
method Trust-Region Iterative Twisted Sequential Monte Carlo (TRI-TSMC) for variance reduction and efficiency.
result Improves primary alignment objectives on text generation tasks.
Nonnegative matrix factorization (NMF), a dimensionality reduction and factor analysis method, is a special case in which factor matrices have low-rank nonnegative constraints. Considering the stochastic learning in NMF, we specifically address the multiplicative update (MU) rule, which is the most popular, but which h…
The ADAM optimizer is exceedingly popular in the deep learning community. Often it works very well, sometimes it doesn't. Why? We interpret ADAM as a combination of two aspects: for each weight, the update direction is determined by the sign of stochastic gradients, whereas the update magnitude is determined by an esti…
The bias-variance tradeoff doesn't always apply in neural networks, contradicting textbook claims.
problem The bias-variance tradeoff is not universally applicable in neural networks, contradicting textbook teachings.
method Extensive experiments and analysis on neural networks, revisiting Geman et al. (1992) experiments.
result Neural networks do not exhibit a bias-variance tradeoff when increasing network width, contradicting textbook claims.
New method makes machine learning approximations unbiased and efficient.
problem Efficient sampling of complex probability distributions.
method Uses autoregressive neural networks with cluster updates and physical symmetries.
result Shows unbiased and low-variance approximations for phase transitions.
Bayesian framework improves variance component estimation in MET data.
problem Inaccurate estimation of variance components in MET data.
method Proposes a Bayesian updating framework using historical data.
result Stabilizes variance component estimation and quantifies uncertainty.
This work analyzes the statistical properties of adaptive gradient methods.
problem Lack of understanding of the statistical properties of adaptive gradient methods.
method Theoretical analyses and experiments on the variance of update magnitudes.
result The variance of update magnitudes is an increasing and bounded function of time, not diverging.
Bayesian filtering optimizes portfolio weights over time with uncertain parameters.
problem Optimizing portfolios over long periods with unknown parameters.
method Bayesian filtering through dynamic linear models for dynamic parameter estimation.
result Bayesian updating improves portfolio performance and is practical.
The paper deals with learning probability distributions of observed data by artificial neural networks. We suggest a so-called gradient conjugate prior (GCP) update appropriate for neural networks, which is a modification of the classical Bayesian update for conjugate priors. We establish a connection between the gradi…
Stochastic gradient descent is the method of choice for large-scale machine learning problems, by virtue of its light complexity per iteration. However, it lags behind its non-stochastic counterparts with respect to the convergence rate, due to high variance introduced by the stochastic updates. The popular Stochastic …
Interpolates between SPG and NeuRD with Capped Implicit Exploration.
problem Combining SPG and NeuRD for better performance in non-stationary environments.
method Introduces Capped Implicit Exploration (CIX) to interpolate between SPG and NeuRD.
result NeuRD-CIX performs well more consistently than NeuRD while retaining NeuRD's advantages.
Stochastic Gradient Descent (SGD) has become one of the most popular optimization methods for training machine learning models on massive datasets. However, SGD suffers from two main drawbacks: (i) The noisy gradient updates have high variance, which slows down convergence as the iterates approach the optimum, and (ii)…
Improved sample complexity for actor-critic algorithms in MDPs.
problem Achieving optimal policies with limited data in reinforcement learning.
method Single-timescale actor-critic with STORM (STOchastic Recursive Momentum) and a sample buffer.
result Optimal sample complexity of O(ε−2) for ε-optimal policies. Stochastic gradient descent updates parameters with summation gradient computed from a random data batch. This summation will lead to unbalanced training process if the data we obtained is unbalanced. To address this issue, this paper takes the error variance and error mean both into consideration. The adaptively adjus…
PES method reduces bias in gradient estimation for unrolled graphs.
problem High variance and bias in gradient estimation for unrolled computation graphs.
method Divide graph into unrolls, apply ES update, accumulate correction terms.
result PES provides unbiased, low-variance gradient estimates.
We introduce novel variants of momentum by incorporating the variance of the stochastic loss function. The variance characterizes the confidence or uncertainty of the local features of the averaged loss surface across the i.i.d. subsets of the training data defined by the mini-batches. We show two applications of the g…
Paper proves SHB convergence with biased gradients and approximate step sizes.
problem Establishing convergence of SHB with biased gradients and approximate step sizes.
method Generalizes SHB convergence conditions for biased gradients, approximate step sizes, and block updating.
result Proves convergence of SHB with new conditions for biased gradients and approximate step sizes.
PLUMAGE improves large model training efficiency and stability.
problem Accelerator memory and networking constraints during large model training.
method Probabilistic Low rank Unbiased Minimum Variance Gradient Estimator (PLUMAGE) that resolves bias and variance issues.
result PLUMAGE reduces training loss by 28% on average across the GLUE benchmark.
A method learns common bias for multiple low-variance tasks without hyper-parameter tuning.
problem Learning common bias for multiple low-variance tasks without manual tuning.
method Two variants of online learning methods (aggressive and lazy) that update bias after each datapoint or at the end of each task.
result Across-tasks regret bound derived for the method, showing faster rates for aggressive variant and standard rates for lazy variant.
Develops a new SPP algorithm with variance reduction for weakly convex optimization.
problem Weakly convex, composite optimization problems.
method Inexact semismooth Newton framework with variance reduction for stochastic proximal point updates.
result Establishes convergence results for the proposed algorithm.
In this paper we study a family of variance reduction methods with randomized batch size---at each step, the algorithm first randomly chooses the batch size and then selects a batch of samples to conduct a variance-reduced stochastic update. We give the linear convergence rate for this framework for composite functions…
Choosing appropriate step sizes is critical for reducing the computational cost of training large-scale neural network models. Mini-batch sub-sampling (MBSS) is often employed for computational tractability. However, MBSS introduces a sampling error, that can manifest as a bias or variance in a line search. This is bec…
We consider 1-qubit mixed quantum state estimation by adaptively updating measurements according to previously obtained outcomes and measurement settings. Updates are determined by the average-variance-optimality (A-optimality) criterion, known in the classical theory of experimental design and applied here to quantum …
New algorithms reduce regret in both stochastic and deterministic environments.
problem Designing algorithms that perform well in both types of MDPs.
method Proposed new environment norms and algorithms with variance-dependent regret bounds.
result First algorithm with simultaneously optimal bounds for both stochastic and deterministic MDPs.
We propose and investigate new complementary methodologies for estimating predictive variance networks in regression neural networks. We derive a locally aware mini-batching scheme that result in sparse robust gradients, and show how to make unbiased weight updates to a variance network. Further, we formulate a heurist…
This work analyzes how often to update the target network in Q-learning.
problem Understanding the optimal frequency of target network updates in Q-learning.
method Formulated target updates as a nested optimization scheme, derived finite-time convergence analysis.
result Optimal target update frequency increases geometrically over time.
Proposes a method to stabilize Black Box Variational Inference using the James-Stein estimator.
problem Stability issues and fine-tuning required in basic Black Box Variational Inference.
method Reframe stochastic gradient ascent as multivariate estimation problem using James-Stein estimator.
result Provides a simpler method with consistent performance in terms of model fit and convergence time.
Multi-step temporal difference (TD) learning is an important approach in reinforcement learning, as it unifies one-step TD learning with Monte Carlo methods in a way where intermediate algorithms can outperform either extreme. They address a bias-variance trade off between reliance on current estimates, which could be …
Unified framework combines views and optimization for better portfolio management.
problem Optimizing portfolio weights with dynamic adjustment based on volatility.
method Dynamic sliding window adjusting horizon, factor estimates, BL posterior returns, and weights over time.
result Outperforms dynamic mean-variance optimization without BL views, providing stronger downside risk control.
With the purpose of examining biased updates in variance-reduced stochastic gradient methods, we introduce SVAG, a SAG/SAGA-like method with adjustable bias. SVAG is analyzed in a cocoercive root-finding setting, a setting which yields the same results as in the usual smooth convex optimization setting for the ordinary…
Paper unifies off-policy learning algorithms and introduces C-trace for better trade-offs.
problem Improving efficiency and scalability in off-policy learning.
method Unified view of off-policy algorithms, considering update variance, fixed-point bias, and contraction rate trade-offs.
result C-trace algorithm demonstrates better trade-offs and state-of-the-art performance.
The policy gradient approach is a flexible and powerful reinforcement learning method particularly for problems with continuous actions such as robot control. A common challenge in this scenario is how to reduce the variance of policy gradient estimates for reliable policy updates. In this paper, we combine the followi…
This paper studies communication efficiency in federated learning by optimizing the sum-rate-distortion function for indirect multiterminal source coding.
problem Indirect multiterminal source coding in federated learning where edge devices send noisy gradients to the server.
method Analyzes the rate region for the quadratic vector Gaussian CEO problem under unbiased estimator and derives an explicit formula for the sum-rate-distortion function.
result Derives an explicit formula for the sum-rate-distortion function in the special case of identical gradients over edge devices.
Drago optimizes DRO problems with faster convergence.
problem Distributionally robust optimization with closed, convex uncertainty sets.
method Primal-dual coupled variance reduction algorithm with cyclic and randomized updates.
result Achieves state-of-the-art linear convergence rate on strongly convex-strongly concave problems.
STORM-PG uses momentum for faster policy gradient updates.
problem Improving policy gradient methods for reinforcement learning.
method Introduces STORM-PG, a SARAH-based algorithm with exponential moving average.
result Achieves O(1/ε3) sample complexity, matching best-known rate. Adaptive OMD reduces variance in learning optimal strategies for imperfect information games.
problem High variance in learning optimal strategies for imperfect information games.
method Fixed sampling approach with locally applied Online Mirror Descent (OMD) algorithm.
result Convergence rate of ildeO(T−1/2) with high probability. AdaQuantFL reduces communication in federated learning by adaptively quantizing model updates.
problem Efficient communication of model updates in federated learning with high-dimensional models and limited bandwidth.
method AdaQuantFL uses adaptive quantization to reduce the number of bits for model updates while maintaining low error floor.
result AdaQuantFL converges in fewer communicated bits compared to fixed quantization levels, with minimal impact on accuracy.
MARS optimizes large model training by reducing variance, outperforming AdamW.
problem Training large models efficiently and scalably.
method Unified optimization framework MARS combining preconditioned gradient updates and variance reduction.
result MARS outperforms AdamW in training GPT-2 models.