Selective Generation for Controllable Language Models
- URL: http://arxiv.org/abs/2307.09254v4
- Date: Mon, 27 Jan 2025 18:45:22 GMT
- Title: Selective Generation for Controllable Language Models
- Authors: Minjae Lee, Kyungmin Kim, Taesoo Kim, Sangdon Park,
- Abstract summary: Trustworthiness of generative language models (GLMs) is crucial in their deployment to critical decision making systems.<n>We propose two selective generation algorithms which control the false discovery rate with respect to the textual entailment relation (FDR-E)<n>$textttSGentextttSup$, a direct modification of the selective prediction, exploits entailment-labeled data, annotated by humans.<n>Since human annotation is costly, we propose a semi-supervised version, $textttSGentextttSemi$, which fully utilizes the un
- Score: 19.909671258499184
- License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
- Abstract: Trustworthiness of generative language models (GLMs) is crucial in their deployment to critical decision making systems. Hence, certified risk control methods such as selective prediction and conformal prediction have been applied to mitigating the hallucination problem in various supervised downstream tasks. However, the lack of appropriate correctness metric hinders applying such principled methods to language generation tasks. In this paper, we circumvent this problem by leveraging the concept of textual entailment to evaluate the correctness of the generated sequence, and propose two selective generation algorithms which control the false discovery rate with respect to the textual entailment relation (FDR-E) with a theoretical guarantee: $\texttt{SGen}^{\texttt{Sup}}$ and $\texttt{SGen}^{\texttt{Semi}}$. $\texttt{SGen}^{\texttt{Sup}}$, a direct modification of the selective prediction, is a supervised learning algorithm which exploits entailment-labeled data, annotated by humans. Since human annotation is costly, we further propose a semi-supervised version, $\texttt{SGen}^{\texttt{Semi}}$, which fully utilizes the unlabeled data by pseudo-labeling, leveraging an entailment set function learned via conformal prediction. Furthermore, $\texttt{SGen}^{\texttt{Semi}}$ enables to use more general class of selection functions, neuro-selection functions, and provides users with an optimal selection function class given multiple candidates. Finally, we demonstrate the efficacy of the $\texttt{SGen}$ family in achieving a desired FDR-E level with comparable selection efficiency to those from baselines on both open and closed source GLMs. Code and datasets are provided at https://github.com/ml-postech/selective-generation.
Related papers
- $\texttt{SEM-CTRL}$: Semantically Controlled Decoding [53.86639808659575]
$texttSEM-CTRL$ is a unified approach that enforces rich context-sensitive constraints and task- and instance-specific semantics directly on an LLM decoder.
texttSEM-CTRL$ allows small pre-trained LLMs to efficiently outperform larger variants and state-of-the-art reasoning models.
arXiv Detail & Related papers (2025-03-03T18:33:46Z) - Robust Information Selection for Hypothesis Testing with Misclassification Penalties [3.3444620077119436]
We study the problem of robust information selection for a Bayesian hypothesis testing / classification task.<n>The goal is to identify the true state of the world from a finite set of hypotheses based on observations from selected information sources.<n>We introduce a novel misclassification penalty framework, which enables non-uniform treatment of different misclassification events.
arXiv Detail & Related papers (2025-02-20T17:05:27Z) - CalibraEval: Calibrating Prediction Distribution to Mitigate Selection Bias in LLMs-as-Judges [21.580762639442913]
We introduce CalibraEval, a novel label-free method for mitigating selection bias during inference.
CalibraEval reformulates debiasing as an optimization task aimed at adjusting observed prediction distributions to align with unbiased prediction distributions.
We show that CalibraEval effectively mitigates selection bias and improves performance compared to existing debiasing methods.
arXiv Detail & Related papers (2024-10-20T13:47:39Z) - OD-Stega: LLM-Based Near-Imperceptible Steganography via Optimized Distributions [7.611860976107124]
We consider coverless steganography where a Large Language Model drives an arithmetic coding decoder to generate stego-texts.
An efficient method should embed secret message bits in as few language tokens as possible, while still keeping the stego-text natural and fluent.
arXiv Detail & Related papers (2024-10-06T01:30:45Z) - Robust Reinforcement Learning from Corrupted Human Feedback [86.17030012828003]
Reinforcement learning from human feedback (RLHF) provides a principled framework for aligning AI systems with human preference data.
We propose a robust RLHF approach -- $R3M$, which models the potentially corrupted preference label as sparse outliers.
Our experiments on robotic control and natural language generation with large language models (LLMs) show that $R3M$ improves robustness of the reward against several types of perturbations to the preference data.
arXiv Detail & Related papers (2024-06-21T18:06:30Z) - Co-training for Low Resource Scientific Natural Language Inference [65.37685198688538]
We propose a novel co-training method that assigns weights based on the training dynamics of the classifiers to the distantly supervised labels.
By assigning importance weights instead of filtering out examples based on an arbitrary threshold on the predicted confidence, we maximize the usage of automatically labeled data.
The proposed method obtains an improvement of 1.5% in Macro F1 over the distant supervision baseline, and substantial improvements over several other strong SSL baselines.
arXiv Detail & Related papers (2024-06-20T18:35:47Z) - Rejection via Learning Density Ratios [50.91522897152437]
Classification with rejection emerges as a learning paradigm which allows models to abstain from making predictions.<n>We propose a different distributional perspective, where we seek to find an idealized data distribution which maximizes a pretrained model's performance.<n>Our framework is tested empirically over clean and noisy datasets.
arXiv Detail & Related papers (2024-05-29T01:32:17Z) - Learnable Item Tokenization for Generative Recommendation [78.30417863309061]
We propose LETTER (a LEarnable Tokenizer for generaTivE Recommendation), which integrates hierarchical semantics, collaborative signals, and code assignment diversity.
LETTER incorporates Residual Quantized VAE for semantic regularization, a contrastive alignment loss for collaborative regularization, and a diversity loss to mitigate code assignment bias.
arXiv Detail & Related papers (2024-05-12T15:49:38Z) - Dirichlet-Based Prediction Calibration for Learning with Noisy Labels [40.78497779769083]
Learning with noisy labels can significantly hinder the generalization performance of deep neural networks (DNNs)
Existing approaches address this issue through loss correction or example selection methods.
We propose the textitDirichlet-based Prediction (DPC) method as a solution.
arXiv Detail & Related papers (2024-01-13T12:33:04Z) - Towards Optimal Statistical Watermarking [95.46650092476372]
We study statistical watermarking by formulating it as a hypothesis testing problem.
Key to our formulation is a coupling of the output tokens and the rejection region.
We characterize the Uniformly Most Powerful (UMP) watermark in the general hypothesis testing setting.
arXiv Detail & Related papers (2023-12-13T06:57:00Z) - RegaVAE: A Retrieval-Augmented Gaussian Mixture Variational Auto-Encoder
for Language Modeling [79.56442336234221]
We introduce RegaVAE, a retrieval-augmented language model built upon the variational auto-encoder (VAE)
It encodes the text corpus into a latent space, capturing current and future information from both source and target text.
Experimental results on various datasets demonstrate significant improvements in text generation quality and hallucination removal.
arXiv Detail & Related papers (2023-10-16T16:42:01Z) - Controlling Federated Learning for Covertness [15.878313629774269]
A learner aims to minimize a function $f$ by repeatedly querying a distributed oracle that provides noisy gradient evaluations.
At the same time, the learner seeks to hide $argmin f$ from a malicious eavesdropper that observes the learner's queries.
This paper considers the problem of textitcovert or textitlearner-private optimization, where the learner has to dynamically choose between learning and obfuscation.
arXiv Detail & Related papers (2023-08-17T07:16:41Z) - An Invariant Learning Characterization of Controlled Text Generation [25.033675230270212]
Controlled generation refers to the problem of creating text that contains stylistic or semantic attributes of interest.
We show that the performance of controlled generation may be poor if the distributions of text in response to user prompts differ from the distribution the predictor was trained on.
arXiv Detail & Related papers (2023-05-31T21:35:08Z) - M-Tuning: Prompt Tuning with Mitigated Label Bias in Open-Set Scenarios [103.6153593636399]
We propose a vision-language prompt tuning method with mitigated label bias (M-Tuning)
It introduces open words from the WordNet to extend the range of words forming the prompt texts from only closed-set label words to more, and thus prompts are tuned in a simulated open-set scenario.
Our method achieves the best performance on datasets with various scales, and extensive ablation studies also validate its effectiveness.
arXiv Detail & Related papers (2023-03-09T09:05:47Z) - AUC-based Selective Classification [5.406386303264086]
We propose a model-agnostic approach to associate a selection function to a given binary classifier.
We provide both theoretical justifications and a novel algorithm, called $AUCross$, to achieve such a goal.
Experiments show that $AUCross$ succeeds in trading-off coverage for AUC, improving over existing selective classification methods targeted at optimizing accuracy.
arXiv Detail & Related papers (2022-10-19T16:29:50Z) - Classifiers are Better Experts for Controllable Text Generation [63.17266060165098]
We show that the proposed method significantly outperforms recent PPLM, GeDi, and DExperts on PPL and sentiment accuracy based on the external classifier of generated texts.
The same time, it is also easier to implement and tune, and has significantly fewer restrictions and requirements.
arXiv Detail & Related papers (2022-05-15T12:58:35Z) - Online Selective Classification with Limited Feedback [82.68009460301585]
We study selective classification in the online learning model, wherein a predictor may abstain from classifying an instance.
Two salient aspects of the setting we consider are that the data may be non-realisable, due to which abstention may be a valid long-term action.
We construct simple versioning-based schemes for any $mu in (0,1],$ that make most $Tmu$ mistakes while incurring smash$tildeO(T1-mu)$ excess abstention against adaptive adversaries.
arXiv Detail & Related papers (2021-10-27T08:00:53Z) - AGGGEN: Ordering and Aggregating while Generating [12.845842212733695]
We present AGGGEN, a data-to-text model which re-introduces two explicit sentence planning stages into neural data-to-text systems.
AGGGEN performs sentence planning at the same time as generating text by learning latent alignments between input representation and target text.
arXiv Detail & Related papers (2021-06-10T08:14:59Z) - Conditioned Text Generation with Transfer for Closed-Domain Dialogue
Systems [65.48663492703557]
We show how to optimally train and control the generation of intent-specific sentences using a conditional variational autoencoder.
We introduce a new protocol called query transfer that allows to leverage a large unlabelled dataset.
arXiv Detail & Related papers (2020-11-03T14:06:10Z) - Improving Text Generation with Student-Forcing Optimal Transport [122.11881937642401]
We propose using optimal transport (OT) to match the sequences generated in training and testing modes.
An extension is also proposed to improve the OT learning, based on the structural and contextual information of the text sequences.
The effectiveness of the proposed method is validated on machine translation, text summarization, and text generation tasks.
arXiv Detail & Related papers (2020-10-12T19:42:25Z) - Rationalizing Text Matching: Learning Sparse Alignments via Optimal
Transport [14.86310501896212]
In this work, we extend this selective rationalization approach to text matching.
The goal is to jointly select and align text pieces, such as tokens or sentences, as a justification for the downstream prediction.
Our approach employs optimal transport (OT) to find a minimal cost alignment between the inputs.
arXiv Detail & Related papers (2020-05-27T01:20:49Z)
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.