Stable cylinders found in hyperbolic groups and curve graphs.
problem Torsionfree hyperbolic groups and curve graphs of surfaces have globally stable cylinders.
method Generalised Sageev's construction to improve fine properties of hyperbolic spaces.
result Proved curve graphs of surfaces admit equivariant quasi-isometric embeddings in finite products of quasitrees.
The paper introduces a Hessian-based method to improve generalization in fine-tuned deep neural networks.
problem Improving generalization in fine-tuned deep neural networks, especially in noisy conditions.
method PAC-Bayesian analysis to identify a Hessian-based distance measure, proving generalization bounds, and developing an algorithm with a generalization error guarantee.
result Hessian-based distance measure correlates well with observed generalization gaps and can match the scale of these gaps in practice.
This work improves online fine-tuning of diffusion models for specific properties.
problem Efficiently fine-tuning diffusion models to maximize specific properties.
method A novel reinforcement learning procedure that efficiently explores feasible samples.
result The method provides a regret guarantee and empirical validation across multiple domains.
Enhances molecular design models by fine-tuning uncertainty-guided VAEs.
problem Fine-tuning pre-trained generative models for specific molecular property optimization.
method Uncertainty-guided fine-tuning of variational autoencoders in an active learning setting.
result Uncertainty-guided fine-tuning improves model performance across multiple molecular properties.
Transformers fine-tuned on synthetic data boost tabular data classification performance.
problem Improving tabular data classification accuracy.
method Fine-tuning ICL-transformers on synthetic datasets with complex decision boundaries.
result Fine-tuned ICL-transformers outperform regular neural networks on real-world datasets.
Improved fine-tuning with regularization and robustness for noisy labels.
problem Fine-tuning pre-trained models on small datasets can lead to overfitting and memorization.
method PAC-Bayes generalization bound analysis, layer-wise regularization, self-label-correction, label-reweighting.
result Improves performance by 1.76% on average for image classification tasks and 0.75% for few-shot classification.
This work classifies SOC loss functions based on their gradient properties.
problem Optimizing noisy systems in stochastic optimal control.
method Grouping loss functions into classes with the same gradient expectation.
result Different loss functions have the same optimization landscape but differ in gradient variance.
Conditional meta-learning improves meta-learning performance in diverse task environments.
problem Meta-learning struggles with tasks that have heterogeneous complexity.
method Conditional meta-learning infers a task-specific meta-parameter vector.
result Conditional meta-learning outperforms standard meta-learning in diverse task environments.
We analyse the fine convergence properties of one parameter families of hyperbolic metrics, on a fixed underlying surface, that move always in a horizontal direction, i.e. orthogonal to the action of diffeomorphisms.
Improves drug properties using a novel LLM and reinforcement learning.
problem Optimizing drug properties while retaining chemical stability.
method Structured Policy Optimization (SPO) for fine-tuning a large language model.
result Enhanced drug properties across multiple target objectives.
Study on singularities in area-minimizing currents, proving unique tangent cones and rectifiability.
problem Understanding singularities in area-minimizing currents.
method Fine excess decay theorems and almost monotonicity of a frequency function.
result Unique tangent cones and countably (m−2)-rectifiable singular set. Understanding the power of depth in feed-forward neural networks is an ongoing challenge in the field of deep learning theory. While current works account for the importance of depth for the expressive power of neural-networks, it remains an open question whether these benefits are exploited during a gradient-based opt…
SMART-FAN-Lasso fine-tunes neural networks for high-dimensional nonparametric regression.
problem Fine-tuning neural networks for high-dimensional nonparametric regression with variable selection.
method Source-model-augmented residual tuning (SMART) framework for neural Lasso.
result SMART-FAN-Lasso achieves statistical acceleration over single-task learning under precise conditions.
The paper studies combinatorics of injective words in the context of Temperley-Lieb algebras.
problem Combinatorial properties of injective words in the context of Temperley-Lieb algebras.
method Investigation of a chain complex of modules over the Temperley-Lieb algebra, focusing on Euler characteristic, homology modules, and Jacobsthal numbers.
result The Euler characteristic of the complex is the n-th Fine number, and the top-dimensional homology module is decomposed in terms of standard Young tableaux.
Simple attack bypasses state-of-the-art DNN watermarking.
problem Protecting DNN models from watermark removal attacks.
method Combining imperceptible pattern embedding and spatial-level transformations for a simple yet effective watermark removal.
result Our attack bypasses state-of-the-art watermarking solutions with high success rates.
The paper characterizes G-ANR spaces and their properties.
problem Characterizing G-ANR spaces and their properties for compact groups. method Proving conditions for a metrizable G-space to be a G-ANR. result Conditions for a metrizable G-space to be a G-ANR are provided. 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) iterations under certain conditions. New model generates images by reversing heat equation, revealing disentanglement.
problem Image generation without considering image structure.
method Stochastically reverses the heat equation to generate images, using variational approximation.
result Emergent disentanglement of overall colour and shape in images.
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.
Pipelined Backpropagation trains large models without batches efficiently.
problem Training large models efficiently on hardware with limited batch sizes.
method Fine-grained Pipelined Backpropagation with Spike Compensation and Linear Weight Prediction.
result Fine-grained Pipelined Backpropagation with a batch size of one matches the accuracy of SGD for multiple networks.
Multirate training speeds up neural network fine-tuning.
problem Efficiently fine-tuning deep neural networks.
method Partitioning neural network parameters into fast and slow parts, updating slowly over longer intervals.
result Significant computational speed-up for transfer learning tasks.
LQF linearizes deep models for better interpretability.
problem Lack of interpretability in deep neural networks.
method Simple modifications to architecture, loss function, and optimization.
result Comparable performance to non-linear fine-tuning, with interpretability.
New measure FTC quantifies how much a ReLU network can fine-tune.
problem Analyzing memorization capacity in fine-tuned neural networks.
method Defined Fine-Tuning Capacity (FTC) for additive fine-tuning of ReLU networks.
result Upper and lower bounds on FTC for 2 and 3-layer ReLU networks.
This paper re-evaluates hyperparameters for fine-tuning pre-trained models.
problem Current hyperparameter settings for fine-tuning are often ad-hoc and fixed.
method Empirical evaluation of learning rate, batch size, and momentum for fine-tuning.
result Optimal hyperparameters are not only dataset-dependent but also sensitive to domain similarity.
Company2Vec creates embeddings from company websites for fine-grained business analytics.
problem Lack of fine-grained company labels for analytics.
method Word2Vec and dimensionality reduction on company website data.
result Semantic company embeddings for various applications.
The paper studies quaternionic Monge-Ampère equations in weighted energy classes.
problem Characterizing the finite energy range of quaternionic Monge-Ampère operator.
method Proving well-definedness and fine property of the operator in weighted energy classes.
result Explicit characterization of the finite energy range of quaternionic Monge-Ampère operator.
New method quantifies uncertainty in fine-tuned LLMs using LoRA ensembles.
problem Uncertainty in fine-tuned LLMs and how to trust their predictions.
method Posterior approximations using low-rank adaptation ensembles.
result Unexpected retention of acquired knowledge during fine-tuning in overfitting regime.
Deep neural networks paved the way for significant improvements in image visual categorization during the last years. However, even though the tasks are highly varying, differing in complexity and difficulty, existing solutions mostly build on the same architectural decisions. This also applies to the selection of acti…
BERT fine-tuning is unstable due to optimization issues, not forgetting or dataset size.
problem Stability of fine-tuning BERT-based models across different random seeds.
method Analysis of BERT, RoBERTa, and ALBERT fine-tuned on GLUE datasets, identifying optimization difficulties as the cause of instability.
result Fine-tuning instability is due to optimization difficulties leading to vanishing gradients, not forgetting or dataset size.
Deep-Lock secures DNN models with secret keys.
problem Preventing unauthorized usage of DNN models.
method Uses S-Boxes and key scheduling to encrypt DNN parameters.
result Ensures correct functioning only with the correct secret key.
Estimates Kaehler metrics' diameter in big cohomology classes.
problem Estimating the diameter of Kaehler metrics in big cohomology classes.
method Proves uniform diameter estimates using integrability conditions and stability properties of complex Monge-Ampere equations.
result Uniform diameter estimates for Kaehler metrics in big cohomology classes.
Fine-tunes GNNs by preserving generative patterns to improve transferability.
problem Vanilla fine-tuning fails due to structural divergence between pre-training and downstream graphs.
method G-Tuning, which reconstructs the generative patterns of the downstream graph using graphon bases.
result G-Tuning achieves an average improvement of 0.5% and 2.6% on in-domain and out-of-domain transfer learning experiments.
Improved code translation by preserving structure with composed fine-tuning.
problem Improving code translation accuracy with unlabeled code outputs.
method Pre-trained denoiser to capture output structure, composed fine-tuning to fine-tune predictor.
result Composed fine-tuning significantly improves generalization over standard fine-tuning.
New findings on hyperbolicity of fine curve graphs and their subgraphs.
problem Investigating hyperbolicity of fine curve graphs and their subgraphs.
method Analyzing large subgraphs of fine curve graphs and computing distances in specific cases.
result Large subgraphs of fine curve graphs contain flats of every finite dimension, indicating they are not hyperbolic.
Homotopy types of curve and arc complexes are studied.
problem Understanding the homotopy types of curve and arc complexes.
method Proving homotopy equivalence and contractibility of complexes.
result Fine curve complex is homotopy equivalent to curve complex, fine arc complex is contractible.
Fine-tuning LLMs improves capability but harms safety, study finds.
problem Balancing capability and safety in LLM fine-tuning.
method Theoretical framework and numerical experiments for two safety-aware fine-tuning strategies.
result Characterization of fundamental limits of safety-capability trade-off in LLM fine-tuning.
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.
This work introduces 'Artificial Entanglement' to understand LLMs' fine-tuning effectiveness.
problem Understanding the effectiveness of parameter-efficient fine-tuning methods for large language models.
method Adopting a quantum-information-inspired perspective, the study measures 'Artificial Entanglement' in neural networks.
result LoRA and FFT induce distinct internal entanglement signatures but not external ones, suggesting a 'no-hair' property.
New method learns optimal variance schedule for diffusion models.
problem Diffusion models' sensitivity to variance schedule.
method Probabilistic conditioning, learning schedule during training.
result Comparable or superior results in super-resolution microscopy and quantitative phase imaging.
Fine-tuning harms in-context learning, but restricting updates to the value matrix improves zero-shot performance.
problem Fine-tuning harms in-context learning, reducing zero-shot performance on unseen tasks.
method Theoretical analysis of linear attention models, identifying conditions for degraded few-shot performance.
result Restricting updates to the value matrix improves zero-shot performance while preserving in-context learning.
Measures consistency of tabular LLM predictions under fine-tuning multiplicity.
problem Conflicting predictions from fine-tuned tabular LLMs.
method Local stability measure in embedding space.
result Probabilistic guarantees on prediction consistency under multiplicity.
Self-play fine-tuning improves diffusion models for text-to-image generation.
problem Plateauing performance of diffusion models after data saturation.
method Self-play fine-tuning (SPIN-Diffusion) using competition among model versions.
result Significantly improved model performance and human preference alignment.
The paper develops a theory linking pretraining and fine-tuning in neural networks.
problem Understanding how initialization choices impact feature learning and generalization in neural networks.
method Analytical theory of diagonal linear networks, deriving generalization error as a function of initialization parameters and task statistics.
result Different initialization choices place networks into four fine-tuning regimes with varying abilities to support feature learning and generalization.
Automorphisms of fine 1-curve graph linked to surface homeomorphisms.
problem Understanding automorphisms of fine 1-curve graphs.
method Isomorphic mapping to surface homeomorphisms.
result Automorphism group is isomorphic to homeomorphism group of a surface.
New method reduces fine-tuning cost for reused models.
problem Repeating fine-tuning costs with outdated foundation models.
method Portable Reward Tuning (PRT) trains a reward model to maximize the same loss function as fine-tuning.
result PRT achieves comparable accuracy to inference-time tuning with less inference cost.
Compact models match or exceed GPT's performance in financial news sentiment analysis.
problem Improving financial sentiment analysis models without large computational costs.
method Fine-tuning non-generative, small-sized models (FinBERT, FinDRoBERTa) on a novel market score database.
result Fine-tuned models outperform GPT-3.5 and GPT-4 in zero-shot learning for financial news sentiment analysis.
Automorphisms of fine graphs for surfaces and tori are studied.
problem Understanding automorphisms of fine graphs for surfaces and tori.
method Extending previous results to tori and discussing smooth versions.
result Automorphism groups of fine graphs for surfaces and tori are naturally isomorphic to homeomorphism groups.
Fine-tuning large language models requires minimal data, making them efficient.
problem Achieving state-of-the-art performance with large language models.
method Using BERT as an example, fine-tuning only the most critical layers of the pre-trained model.
result Fine-tuned models are close in parameter space to the pre-trained model, with many good solutions found in sparsified versions.