Continuous Language Model Interpolation for Dynamic and Controllable Text Generation

• URL: http://arxiv.org/abs/2404.07117v1
• Date: Wed, 10 Apr 2024 15:55:07 GMT
• Title: Continuous Language Model Interpolation for Dynamic and Controllable Text Generation
• Authors: Sara Kangaslahti, David Alvarez-Melis,
• Abstract summary: We focus on the challenging case where the model must dynamically adapt to diverse -- and often changing -- user preferences. We leverage adaptation methods based on linear weight, casting them as continuous multi-domain interpolators. We show that varying the weights yields predictable and consistent change in the model outputs.
• Score: 7.535219325248997
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: As large language models (LLMs) have gained popularity for a variety of use cases, making them adaptable and controllable has become increasingly important, especially for user-facing applications. While the existing literature on LLM adaptation primarily focuses on finding a model (or models) that optimizes a single predefined objective, here we focus on the challenging case where the model must dynamically adapt to diverse -- and often changing -- user preferences. For this, we leverage adaptation methods based on linear weight interpolation, casting them as continuous multi-domain interpolators that produce models with specific prescribed generation characteristics on-the-fly. Specifically, we use low-rank updates to fine-tune a base model to various different domains, yielding a set of anchor models with distinct generation profiles. Then, we use the weight updates of these anchor models to parametrize the entire (infinite) class of models contained within their convex hull. We empirically show that varying the interpolation weights yields predictable and consistent change in the model outputs with respect to all of the controlled attributes. We find that there is little entanglement between most attributes and identify and discuss the pairs of attributes for which this is not the case. Our results suggest that linearly interpolating between the weights of fine-tuned models facilitates predictable, fine-grained control of model outputs with respect to multiple stylistic characteristics simultaneously.

Related papers

• You Only Merge Once: Learning the Pareto Set of Preference-Aware Model Merging [11.186194228460273]
  We propose preference-aware model merging in which the performance of the merged model on each base model's task is treated as an objective. In only one merging process, the proposed parameter-efficient structure can generate the whole set of merged models. We show that the proposed preference-aware model merging can obtain a diverse set of trade-off models and outperforms state-of-the-art model merging baselines.
  arXiv  Detail & Related papers  (2024-08-22T03:41:14Z)
• SMILE: Zero-Shot Sparse Mixture of Low-Rank Experts Construction From Pre-Trained Foundation Models [85.67096251281191]
  We present an innovative approach to model fusion called zero-shot Sparse MIxture of Low-rank Experts (SMILE) construction. SMILE allows for the upscaling of source models into an MoE model without extra data or further training. We conduct extensive experiments across diverse scenarios, such as image classification and text generation tasks, using full fine-tuning and LoRA fine-tuning.
  arXiv  Detail & Related papers  (2024-08-19T17:32:15Z)
• EMR-Merging: Tuning-Free High-Performance Model Merging [55.03509900949149]
  We show that Elect, Mask & Rescale-Merging (EMR-Merging) shows outstanding performance compared to existing merging methods. EMR-Merging is tuning-free, thus requiring no data availability or any additional training while showing impressive performance.
  arXiv  Detail & Related papers  (2024-05-23T05:25:45Z)
• The Interpolating Information Criterion for Overparameterized Models [49.283527214211446]
  We show that the Interpolating Information Criterion is a measure of model quality that naturally incorporates the choice of prior into the model selection. Our new information criterion accounts for prior misspecification, geometric and spectral properties of the model, and is numerically consistent with known empirical and theoretical behavior.
  arXiv  Detail & Related papers  (2023-07-15T12:09:54Z)
• Fast Adaptation with Bradley-Terry Preference Models in Text-To-Image Classification and Generation [0.0]
  We leverage the Bradley-Terry preference model to develop a fast adaptation method that efficiently fine-tunes the original model. Extensive evidence of the capabilities of this framework is provided through experiments in different domains related to multimodal text and image understanding.
  arXiv  Detail & Related papers  (2023-07-15T07:53:12Z)
• Dataless Knowledge Fusion by Merging Weights of Language Models [51.8162883997512]
  Fine-tuning pre-trained language models has become the prevalent paradigm for building downstream NLP models. This creates a barrier to fusing knowledge across individual models to yield a better single model. We propose a dataless knowledge fusion method that merges models in their parameter space.
  arXiv  Detail & Related papers  (2022-12-19T20:46:43Z)
• CAMERO: Consistency Regularized Ensemble of Perturbed Language Models with Weight Sharing [83.63107444454938]
  We propose a consistency-regularized ensemble learning approach based on perturbed models, named CAMERO. Specifically, we share the weights of bottom layers across all models and apply different perturbations to the hidden representations for different models, which can effectively promote the model diversity. Our experiments using large language models demonstrate that CAMERO significantly improves the generalization performance of the ensemble model.
  arXiv  Detail & Related papers  (2022-04-13T19:54:51Z)
• Model Compression for Domain Adaptation through Causal Effect Estimation [20.842938440720303]
  ATE-guided Model Compression scheme (AMoC) generates many model candidates, differing by the model components that were removed. Then, we select the best candidate through a stepwise regression model that utilizes the ATE to predict the expected performance on the target domain. AMoC outperforms strong baselines on 46 of 60 domain pairs across two text classification tasks, with an average improvement of more than 3% in F1 above the strongest baseline.
  arXiv  Detail & Related papers  (2021-01-18T14:18:02Z)
• Improving the Reconstruction of Disentangled Representation Learners via Multi-Stage Modeling [54.94763543386523]
  Current autoencoder-based disentangled representation learning methods achieve disentanglement by penalizing the ( aggregate) posterior to encourage statistical independence of the latent factors. We present a novel multi-stage modeling approach where the disentangled factors are first learned using a penalty-based disentangled representation learning method. Then, the low-quality reconstruction is improved with another deep generative model that is trained to model the missing correlated latent variables.
  arXiv  Detail & Related papers  (2020-10-25T18:51:15Z)
• Reinforcement Learning based dynamic weighing of Ensemble Models for Time Series Forecasting [0.8399688944263843]
  It is known that if models selected for data modelling are distinct (linear/non-linear, static/dynamic) and independent (minimally correlated) models, the accuracy of the predictions is improved. Various approaches suggested in the literature to weigh the ensemble models use a static set of weights. To address this issue, a Reinforcement Learning (RL) approach to dynamically assign and update weights of each of the models at different time instants.
  arXiv  Detail & Related papers  (2020-08-20T10:40:42Z)

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.