Research
On-device research index

arXiv research

A locally-built, LLM-digested index of recent arXiv papers in quant finance, geometry/topology, and statistical ML — keyword search served straight from SQLite on this machine.

168,657 papers · 148 categories

Trend · papers per month

86172257343 · Jun 202019922001200920172026
48 results for weight tuning

This paper optimizes binary linear classifiers by tuning their weight vectors.

problem Optimizing the weight vector of binary linear classifiers for better performance.
method Parameterization of the discriminant through a scalar to control trade-offs between informative and noisy terms.
result Weight vector tuning compensates for non-optimal native hyperparameters, improving classification performance.

Optimizes sparse fine-tuning for privacy in neural networks.

problem Performance gap between DP-SGD and non-private fine-tuning.
method Optimization-based approach using private gradient information for selecting trainable weights.
result Our selection method leads to better prediction accuracy compared to existing approaches.

Weight Squeezing transfers knowledge from large models to smaller ones, improving performance and speed.

problem Transfer learning and model compression for faster and more efficient training.
method Reparameterization of weights from a large model to a smaller one.
result Weight Squeezing outperforms other methods on GLUE benchmark with faster training.

Generative approach speeds hyperparameter tuning for machine learning models.

problem Computational infeasibility of cross-validation and difficulty of fully Bayesian hyper-parameter learning.
method Combines optimization-based approximations and amortization techniques.
result Rapid evaluation of hyper-parameters over grids or ranges, supporting predictive tuning and uncertainty quantification.

Unified approach to linear regression using covariance fitting for optimal weights.

problem Finding optimal weights for linear regression models when weights are unknown.
method Covariance fitting SPICE-methodology to obtain data-adaptive weights.
result Tuned versions of known regularized estimators are unified under a common approach.

We investigate approaches to regularisation during fine-tuning of deep neural networks. First we provide a neural network generalisation bound based on Rademacher complexity that uses the distance the weights have moved from their initial values. This bound has no direct dependence on the number of weights and compares…

2020-02-19abs ↗pdf ↗

Paper proves multiplicative weight updates can train neural networks without learning rate tuning.

problem Vanishing and exploding gradients in gradient descent for compositional functions.
method Proves descent lemma for compositional functions using multiplicative weight updates and derives Madam optimizer.
result Madam optimizer trains state-of-the-art neural networks without learning rate tuning.

Researchers develop method to protect against 'weight poisoning' attacks on pre-trained models.

problem The security threat of downloading untrusted pre-trained weights that can be manipulated after fine-tuning.
method RIPPLe regularization method and Embedding Surgery initialization procedure.
result Demonstrated that weight poisoning attacks are possible even with limited knowledge of the dataset and fine-tuning procedure.

New framework for data-driven hyperparameter tuning with structured loss.

problem Statistical foundations for multi-dimensional hyperparameter tuning remain limited.
method General framework using real algebraic geometry for semi-algebraic function classes.
result First general guarantees for multi-dimensional hyperparameter tuning.

LoRA fine-tuning explained with gradient dynamics for low-rank perturbations.

problem Understanding why gradient descent converges to useful low-rank perturbations in LoRA fine-tuning.
method Generalized student-teacher setting with i.i.d. samples and online gradient descent.
result Gradient descent converges to the teacher model in dkO(1)dk^{O(1)} iterations under certain conditions.

We propose an algorithm for inexpensive gradient-based hyperparameter optimization that combines the implicit function theorem (IFT) with efficient inverse Hessian approximations. We present results about the relationship between the IFT and differentiating through optimization, motivating our algorithm. We use the pro…

2019-11-06abs ↗pdf ↗

In this paper, we present a novel approach for fine-tuning a decoder-side neural network in the context of image compression, such that the weight-updates are better compressible. At encoder side, we fine-tune a pre-trained artifact removal network on target data by using a compression objective applied on the weight-u…

2019-05-10abs ↗pdf ↗

This paper explains how batch normalization auto-tunes the regularization parameter based on data statistics.

problem Batch normalization accelerates deep learning training but the exact relationship to regularization is unclear.
method Theoretical analysis and empirical validation of batch normalization's role in auto-tuning the regularization parameter.
result Batch normalization auto-tunes the regularization parameter based on data statistics.

Optimizes pruning masks for neural networks using probabilistic fine-tuning and PAC-Bayes bounds.

problem Improving neural network performance through adaptive pruning of weights.
method Optimizes stochastic pruning masks by minimizing expected loss, considering data-adaptive regularization and feature alignment.
result Probabilistic fine-tuning leads to improved test error over baseline methods in neural networks.

Efficiently tunes hyperparameters with dynamic accuracy method.

problem Optimizing machine learning hyperparameters with inexact evaluations.
method Dynamic accuracy derivative-free optimization for hyperparameter tuning.
result Demonstrates robust and efficient hyperparameter tuning compared to fixed accuracy methods.

A new meta-learning framework that assigns weights to source tasks based on target samples.

problem Learning initialization for target tasks with limited labeled examples.
method A general framework that assigns weights to the loss of different source tasks, which can depend on the target samples. Provides upper bounds and develops a learning algorithm based on minimizing the error bound with respect to an empirical IPM.
result Empirically, the weighted meta-learning algorithm finds better initializations than uniformly-weighted meta-learning algorithms.

EVA adapts LoRA for faster, more efficient fine-tuning.

problem Fast and efficient fine-tuning of large models for specific tasks.
method EVA uses directions capturing most activation variance for initialization, maximizing gradient signal and reducing parameters.
result EVA achieves faster convergence and higher average scores across tasks, reducing parameters.

Proposes a new RL method to fine-tune flow-based models with arbitrary rewards.

problem Challenges in fine-tuning continuous flow-based generative models with arbitrary reward functions.
method Online Reward-Weighted Conditional Flow Matching with Wasserstein-2 Regularization (ORW-CFM-W2)
result Achieves optimal policy convergence with controllable trade-offs between reward maximization and diversity preservation.

This tutorial reviews RL-based methods for optimizing diffusion models to maximize specific metrics.

problem Optimizing diffusion models to generate samples that maximize specific metrics in practical applications.
method Various RL algorithms including PPO, differentiable optimization, reward-weighted MLE, value-weighted sampling, and path consistency learning.
result Exploration of strengths and limitations of RL-based fine-tuning algorithms and their benefits compared to non-RL-based approaches.

A new method aligns source and target distributions by tuning their weights.

problem Domain adaptation on unlabeled target datasets using labeled source datasets.
method Weighted Joint Distribution Optimal Transport (WJDOT) method that finds alignment between source and target distributions and re-weighting of source distributions.
result Achieves state-of-the-art performance on simulated and real-life datasets.

Neural networks with learned biases can approximate any function.

problem Whether neural networks with only learned biases can approximate any continuous function.
method Theoretical and numerical analysis of random weights and learned biases in neural networks.
result Feedforward and recurrent neural networks with random weights can approximate any continuous function and dynamical systems.

Residual Network (ResNet) is the state-of-the-art architecture that realizes successful training of really deep neural network. It is also known that good weight initialization of neural network avoids problem of vanishing/exploding gradients. In this paper, simplified models of ResNets are analyzed. We argue that good…

2017-09-09abs ↗pdf ↗

InfoSFT improves LLMs by focusing on informative, medium-confidence tokens.

problem Overfitting to unlikely samples and degradation of prior capabilities in SFT.
method InfoSFT uses a principled weighting scheme to concentrate learning signals on medium-confidence tokens.
result InfoSFT improves generalization and preserves pre-existing capabilities over vanilla SFT and likelihood-weighted baselines.

New method adapts neural networks without losing prior knowledge.

problem Understanding and enabling flexible adaptation of neural networks.
method Differential geometry framework, functionally invariant paths (FIP).
result Achieves comparable state-of-the-art performance on continual learning and sparsification tasks.

The neural plausibility of backpropagation has long been disputed, primarily for its use of non-local weight transport - the biologically dubious requirement that one neuron instantaneously measure the synaptic weights of another. Until recently, attempts to create local learning rules that avoid weight transport hav…

2020-02-28abs ↗pdf ↗

DTM improves dLLM fine-tuning stability and performance.

problem Intractable sequence-level marginal likelihoods for masked diffusion models.
method Discrete Tilt Matching (DTM) recasts dLLM fine-tuning as state-level matching of local unmasking posteriors under reward tilting.
result DTM yields strong gains on Sudoku and Countdown while remaining competitive on MATH500 and GSM8K.

Fine-tuning neural networks to guarantee performance on specific examples can also introduce incorrect inputs.

problem Ensuring reliable performance of neural networks on specific examples.
method Using SMT solvers to fine-tune ReLU neural networks to guarantee outcomes on a finite set of particular examples.
result Fine-tuning can introduce incorrect inputs that trigger unexpected performance.

New methods improve LLM preference optimization by intelligently weighting multiple reference models.

problem Improving LLM preference optimization with multiple reference models.
method Introducing four new weighting strategies for multiple-reference preference optimization.
result All four new weighting strategies outperform current methods on preference accuracy.

The study introduces anytime learning schedules for large language models without fixed horizons.

problem Training large language models without knowing the total training horizon.
method Theoretical analysis and weight averaging to create anytime learning schedules.
result Theoretical and empirical evidence shows that weight averaging with simple step sizes can achieve comparable final loss to well-tuned cosine schedules.

The paper identifies patterns in language model weights used for memorizing paragraphs.

problem Locating the specific mechanisms and weights used by language models to memorize paragraphs.
method Examined gradients and attention patterns in language models to identify memorized paragraphs.
result Gradients of memorized paragraphs have a distinguishable spatial pattern, and localized attention heads are involved in paragraph memorization.

Optimizer memory affects learning rate sensitivity in shuffle order, impacting fine-tuning noise.

problem Optimizer memory affects the learning rate sensitivity in shuffle order, leading to fine-tuning noise.
method Isolated the mechanism of fixed-clock optimizer memory affecting the learning rate sensitivity in shuffle order, deriving a fit-free way to size the noise.
result Fixed-clock optimizers like AdamW produce a larger first-order noise channel compared to memoryless optimizers, affecting fine-tuning comparisons.

Jointly tuning ensemble models improves performance and uncertainty calibration.

problem Improving both predictive performance and uncertainty calibration in deep ensembles.
method Investigated the impact of jointly tuning weight decay, temperature scaling, and early stopping.
result Jointly tuning ensemble models generally matches or improves performance, with significant variation across tasks.