Posterior Control of Blackbox Generation

• URL: http://arxiv.org/abs/2005.04560v1
• Date: Sun, 10 May 2020 03:22:45 GMT
• Title: Posterior Control of Blackbox Generation
• Authors: Xiang Lisa Li and Alexander M. Rush
• Abstract summary: We consider augmenting neural generation models with discrete control states learned through a structured latent-variable approach. We find that this method improves over standard benchmarks, while also providing fine-grained control.
• Score: 126.33511630879713
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Text generation often requires high-precision output that obeys task-specific rules. This fine-grained control is difficult to enforce with off-the-shelf deep learning models. In this work, we consider augmenting neural generation models with discrete control states learned through a structured latent-variable approach. Under this formulation, task-specific knowledge can be encoded through a range of rich, posterior constraints that are effectively trained into the model. This approach allows users to ground internal model decisions based on prior knowledge, without sacrificing the representational power of neural generative models. Experiments consider applications of this approach for text generation. We find that this method improves over standard benchmarks, while also providing fine-grained control.

Related papers

• Plug-and-Play Controllable Generation for Discrete Masked Models [27.416952690340903]
  This article makes discrete masked models for the generative modeling of discrete data controllable. We propose a novel plug-and-play framework based on importance sampling that bypasses the need for training a conditional score. Our framework is agnostic to the choice of control criteria, requires no gradient information, and is well-suited for tasks such as posterior sampling, Bayesian inverse problems, and constrained generation.
  arXiv  Detail & Related papers  (2024-10-03T02:00:40Z)
• disco: a toolkit for Distributional Control of Generative Models [4.662591864499645]
  We present disco, an open-source Python library that brings distributional control techniques to the broader public. Despite their potential, the widespread adoption of these techniques has been hindered by the difficulty in adapting complex, disconnected code.
  arXiv  Detail & Related papers  (2023-03-08T18:58:52Z)
• Controllable Text Generation with Neurally-Decomposed Oracle [91.18959622763055]
  We propose a framework to control auto-regressive generation models with NeurAlly-Decomposed Oracle (NADO) We present a closed-form optimal solution to incorporate the token-level guidance into the base model for controllable generation.
  arXiv  Detail & Related papers  (2022-05-27T20:17:53Z)
• Deep Latent-Variable Models for Text Generation [7.119436003155924]
  Deep neural network-based end-to-end architectures have been widely adopted. End-to-end approach conflates all sub-modules, which used to be designed by complex handcrafted rules, into a holistic encode-decode architecture. This dissertation presents how deep latent-variable models can improve over the standard encoder-decoder model for text generation.
  arXiv  Detail & Related papers  (2022-03-03T23:06:39Z)
• Is Disentanglement enough? On Latent Representations for Controllable Music Generation [78.8942067357231]
  In the absence of a strong generative decoder, disentanglement does not necessarily imply controllability. The structure of the latent space with respect to the VAE-decoder plays an important role in boosting the ability of a generative model to manipulate different attributes.
  arXiv  Detail & Related papers  (2021-08-01T18:37:43Z)
• Neural Rule-Execution Tracking Machine For Transformer-Based Text Generation [43.71069101841354]
  Sequence-to-Sequence (S2S) neural text generation models have exhibited compelling performance on various natural language generation tasks. However, the black-box nature of these models limits their application in tasks where specific rules need to be executed. We propose a novel module named Neural Rule-Execution Tracking Machine that can be equipped into various transformer-based generators to leverage multiple rules simultaneously.
  arXiv  Detail & Related papers  (2021-07-27T20:41:05Z)
• Zero-Shot Controlled Generation with Encoder-Decoder Transformers [8.506451605253517]
  We propose novel approaches for controlling encoder-decoder transformer-based NLG models in zero-shot. We show that not only are these NLG models robust to such manipulations, but also their behavior could be controlled without an impact on their generation performance.
  arXiv  Detail & Related papers  (2021-06-11T14:07:19Z)
• NeuroLogic Decoding: (Un)supervised Neural Text Generation with Predicate Logic Constraints [75.66980495245926]
  Conditional text generation often requires lexical constraints, i.e., which words should or shouldn't be included in the output text. We propose NeuroLogic Decoding, a simple yet effective algorithm that enables neural language models -- supervised or not -- to generate fluent text. Our results suggest the limit of large-scale neural networks for fine-grained controllable generation and the promise of inference-time algorithms.
  arXiv  Detail & Related papers  (2020-10-24T11:55:22Z)
• Rewriting a Deep Generative Model [56.91974064348137]
  We introduce a new problem setting: manipulation of specific rules encoded by a deep generative model. We propose a formulation in which the desired rule is changed by manipulating a layer of a deep network as a linear associative memory. We present a user interface to enable users to interactively change the rules of a generative model to achieve desired effects.
  arXiv  Detail & Related papers  (2020-07-30T17:58:16Z)
• A Controllable Model of Grounded Response Generation [122.7121624884747]
  Current end-to-end neural conversation models inherently lack the flexibility to impose semantic control in the response generation process. We propose a framework that we call controllable grounded response generation (CGRG) We show that using this framework, a transformer based model with a novel inductive attention mechanism, trained on a conversation-like Reddit dataset, outperforms strong generation baselines.
  arXiv  Detail & Related papers  (2020-05-01T21:22:08Z)

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.