Hyperbolic Fine-tuning for Large Language Models

• URL: http://arxiv.org/abs/2410.04010v1
• Date: Sat, 5 Oct 2024 02:58:25 GMT
• Title: Hyperbolic Fine-tuning for Large Language Models
• Authors: Menglin Yang, Aosong Feng, Bo Xiong, Jihong Liu, Irwin King, Rex Ying,
• Abstract summary: This study investigates the non-Euclidean characteristics of large language models (LLMs) We show that token embeddings exhibit a high degree of hyperbolicity, indicating a latent tree-like structure in the embedding space. We introduce a new method called hyperbolic low-rank efficient fine-tuning, HypLoRA, that performs low-rank adaptation directly on the hyperbolic manifold.
• Score: 56.54715487997674
• License: http://creativecommons.org/licenses/by-nc-sa/4.0/
• Abstract: Large language models (LLMs) have demonstrated remarkable performance on various tasks. However, it remains an open question whether the default Euclidean space is the most suitable choice for embedding tokens in LLMs. In this study, we first investigate the non-Euclidean characteristics of LLMs. Our findings reveal that token frequency follows a power-law distribution, with high-frequency tokens clustering near the origin and low-frequency tokens positioned farther away. Additionally, token embeddings exhibit a high degree of hyperbolicity, indicating a latent tree-like structure in the embedding space. Building on the observation, we propose to efficiently fine-tune LLMs in hyperbolic space to better exploit the underlying complex structures. However, we found that this fine-tuning in hyperbolic space cannot be achieved with naive application of exponential and logarithmic maps, when the embedding and weight matrices both reside in Euclidean space. To address this technique issue, we introduce a new method called hyperbolic low-rank efficient fine-tuning, HypLoRA, that performs low-rank adaptation directly on the hyperbolic manifold, avoiding the cancellation effect caused by the exponential and logarithmic maps, thus preserving the hyperbolic modeling capabilities. Through extensive experiments, we demonstrate that HypLoRA significantly enhances the performance of LLMs on reasoning tasks, particularly for complex reasoning problems. In particular, HypLoRA improves the performance in the complex AQuA dataset by up to 13.0%, showcasing its effectiveness in handling complex reasoning challenges

Related papers

• Is Parameter Collision Hindering Continual Learning in LLMs? [50.57658782050275]
  Large Language Models (LLMs) often suffer from catastrophic forgetting when learning multiple tasks sequentially. We show that building non-collision parameters is a more critical interdependence factor in addressing CL challenges. We propose Non-collision Low-Rank Adaptation (N-LoRA), a simple yet effective approach leveraging low collision rates to enhance CL in LLMs.
  arXiv  Detail & Related papers  (2024-10-14T05:54:11Z)
• SubZero: Random Subspace Zeroth-Order Optimization for Memory-Efficient LLM Fine-Tuning [66.27334633749734]
  As language models grow in size, memory demands for backpropagation increase. Zeroth-order (ZOZO) optimization methods offer a memory-efficient alternative. We show that SubZero enhances fine-tuning and achieves faster results compared to standard ZOZO approaches.
  arXiv  Detail & Related papers  (2024-10-11T17:01:43Z)
• Fast hyperboloid decision tree algorithms [0.6656737591902598]
  We present hyperDT, a novel extension of decision tree algorithms into hyperbolic space. Our approach is conceptually straightforward and maintains constant-time decision complexity. Building upon hyperDT we introduce hyperRF, a hyperbolic random forest model.
  arXiv  Detail & Related papers  (2023-10-20T22:31:10Z)
• The Numerical Stability of Hyperbolic Representation Learning [36.32817250000654]
  We analyze the limitations of two popular models for the hyperbolic space, namely, the Poincar'e ball and the Lorentz model. We extend this Euclidean parametrization to hyperbolic hyperplanes and exhibit its ability to improve the performance of hyperbolic SVM.
  arXiv  Detail & Related papers  (2022-10-31T22:51:59Z)
• HRCF: Enhancing Collaborative Filtering via Hyperbolic Geometric Regularization [52.369435664689995]
  We introduce a textitHyperbolic Regularization powered Collaborative Filtering (HRCF) and design a geometric-aware hyperbolic regularizer. Specifically, the proposal boosts optimization procedure via the root alignment and origin-aware penalty. Our proposal is able to tackle the over-smoothing problem caused by hyperbolic aggregation and also brings the models a better discriminative ability.
  arXiv  Detail & Related papers  (2022-04-18T06:11:44Z)
• Provably Accurate and Scalable Linear Classifiers in Hyperbolic Spaces [39.71927912296049]
  We propose a unified framework for learning scalable and simple hyperbolic linear classifiers. The gist of our approach is to focus on Poincar'e ball models and formulate the classification problems using tangent space formalisms. The excellent performance of the Poincar'e second-order and strategic perceptrons shows that the proposed framework can be extended to general machine learning problems in hyperbolic spaces.
  arXiv  Detail & Related papers  (2022-03-07T21:36:21Z)
• Highly Scalable and Provably Accurate Classification in Poincare Balls [40.82908295137667]
  We establish a unified framework for learning scalable and simple hyperbolic linear classifiers with provable performance guarantees. Our results include a new hyperbolic and second-order perceptron algorithm as well as an efficient and highly accurate convex optimization setup for hyperbolic support vector machine classifiers. Their performance accuracies on synthetic data sets comprising millions of points, as well as on complex real-world data sets such as single-cell RNA-seq expression measurements, CIFAR10, Fashion-MNIST and mini-ImageNet.
  arXiv  Detail & Related papers  (2021-09-08T16:59:39Z)
• Augmented Sliced Wasserstein Distances [55.028065567756066]
  We propose a new family of distance metrics, called augmented sliced Wasserstein distances (ASWDs) ASWDs are constructed by first mapping samples to higher-dimensional hypersurfaces parameterized by neural networks. Numerical results demonstrate that the ASWD significantly outperforms other Wasserstein variants for both synthetic and real-world problems.
  arXiv  Detail & Related papers  (2020-06-15T23:00:08Z)
• Robust Large-Margin Learning in Hyperbolic Space [64.42251583239347]
  We present the first theoretical guarantees for learning a classifier in hyperbolic rather than Euclidean space. We provide an algorithm to efficiently learn a large-margin hyperplane, relying on the careful injection of adversarial examples. We prove that for hierarchical data that embeds well into hyperbolic space, the low embedding dimension ensures superior guarantees.
  arXiv  Detail & Related papers  (2020-04-11T19:11:30Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.