许多文献表明，基于及时的学习是使用大型预训练的语言模型的有效方法。最近的作品还展示了通过插入适当的提示来指导聊天机器人输出的可能性。基于梯度的方法通常用于扰动提示。但是，某些语言模型甚至无法为公众提供。在这项工作中，我们首先探讨了提示和加强学习（RL）与转向模型的生成的组合，而无需访问任何模型的参数。其次，为了减少培训工作并增强对看不见的任务的普遍性，我们应用多任务学习以使模型学会更好地对新任务进行推广。实验结果表明，我们提出的方法可以成功控制几个最新的（SOTA）对话模型，而无需访问其参数。此外，该模型证明了与基线模型更少的步骤快速适应看不见的任务的强大能力。

Controllable Text Generation (CTG) is emerging area in the field of natural language generation (NLG). It is regarded as crucial for the development of advanced text generation technologies that are more natural and better meet the specific constraints in practical applications. In recent years, methods using large-scale pre-trained language models (PLMs), in particular the widely used transformer-based PLMs, have become a new paradigm of NLG, allowing generation of more diverse and fluent text. However, due to the lower level of interpretability of deep neural networks, the controllability of these methods need to be guaranteed. To this end, controllable text generation using transformer-based PLMs has become a rapidly growing yet challenging new research hotspot. A diverse range of approaches have emerged in the recent 3-4 years, targeting different CTG tasks which may require different types of controlled constraints. In this paper, we present a systematic critical review on the common tasks, main approaches and evaluation methods in this area. Finally, we discuss the challenges that the field is facing, and put forward various promising future directions. To the best of our knowledge, this is the first survey paper to summarize CTG techniques from the perspective of PLMs. We hope it can help researchers in related fields to quickly track the academic frontier, providing them with a landscape of the area and a roadmap for future research.

预先接受的语言模型（PLM）在神经对话建模中标志着巨大的飞跃。虽然PLMS在大型文本语料库上进行预先培训，但通常在具有特定领域知识和对话风格的稀缺对话数据上进行微调。然而，在大型预先训练模型中充分利用现有知识的同时定制语言模型仍然是一个挑战。在本文中，我们提出了一种预先接受训练的对话建模的新方法，将对话生成问题作为一个快速学习任务。而不是在有限的对话数据上进行微调，我们的方法，DialogPrompt学习针对对话背景优化的连续提示嵌入，从而从大型预训练模型中促进了知识。为了鼓励模型更好地利用提示嵌入，提示编码器被设计为在输入对话框上下文中的条件。流行对话数据集的实验表明，我们的方法显着优于微调基线和通用及时学习方法。此外，人类评估强烈支持对DialialPrompt的优越性在响应生成质量方面。

在本文中，我们介绍了基于大型预训练的语言模型（PLM）pangu-alpha（Zeng等，2021）的中国预训练的开放域对话生成模型。与其他对大量对话数据进行培训的预训练的对话模型不同，我们旨在通过继承PLM的有价值的语言能力和知识来构建强大的对话模型，并以相对较少的数据和计算成本构建强大的对话模型。为此，我们训练大型PLM Pangu-Alpha的Pangu-bot，该机器人已被证明在各种中国自然语言任务上表现出色。我们研究了pangu-bot产生的响应的不同方面，包括响应质量，知识和安全性。我们表明，Pangu-Bot优于最先进的中国对话系统（CDIALGPT（Wang等，2020），Eva（Zhou等，2021），EVA2.0（Gu等，2022）） W.R.T.以上三个方面。我们还证明，可以轻松地部署pangu-bot，以在没有进一步训练的情况下产生情感反应。在整个经验分析中，我们还指出，Pangu-bot响应质量，知识正确性和安全性仍然远非完美，进一步的探索对于建立可靠且智能的对话系统是必不可少的。我们的型号和代码将在https://github.com/huawei-noah/pretretaining-language-model/tree/master/master/pangu-bot上提供。

Steering language generation towards objectives or away from undesired content has been a long-standing goal in utilizing language models (LM). Recent work has demonstrated reinforcement learning and weighted decoding as effective approaches to achieve a higher level of language control and quality with pros and cons. In this work, we propose a novel critic decoding method for controlled language generation (CriticControl) that combines the strengths of reinforcement learning and weighted decoding. Specifically, we adopt the actor-critic framework to train an LM-steering critic from non-differentiable reward models. And similar to weighted decoding, our method freezes the language model and manipulates the output token distribution using called critic, improving training efficiency and stability. Evaluation of our method on three controlled generation tasks, namely topic control, sentiment control, and detoxification, shows that our approach generates more coherent and well-controlled texts than previous methods. In addition, CriticControl demonstrates superior generalization ability in zero-shot settings. Human evaluation studies also corroborate our findings.

没有一致响应的对话系统并不令人着迷。在这项研究中，我们建立了一个对话系统，可以根据给定的角色设置（角色）响应以带来一致性。考虑到语言模型迅速增加的趋势，我们提出了一种使用迅速调整的方法，该方法在预训练的大规模语言模型上使用了低学习成本。英语和日语中自动和手动评估的结果表明，可以使用比微调更少的计算资源来构建具有更自然和个性化响应的对话系统。

近年来，对话系统引起了学术界和工业的重要兴趣。特别是开放式对话系统的纪律，又名聊天，已经获得了很大的势头。然而，困扰研究人员的长期挑战是缺乏有效的自动评估指标，这导致目前研究中的障碍。评估开放式对话模型表现的常见做法涉及对最终部署模型的广泛人类评估，这是时间和成本密集的。此外，最近建立开放式聊天聊天的趋势涉及具有大量社交媒体对话数据的预训练对话模型。但是，社交媒体对话中包含的信息可能是令人反感的和不合适的。不分青红皂白种的使用可能导致不敏感和有毒的生成模型。本文介绍了对话系统技术挑战10（DSTC10）的轨道5获得的数据，基线和结果。

基于变压器的语言模型能够生成流利的文本，并在各种自然语言生成任务中有效地适应。但是，已证明在大型未标记的网络文本语料库中鉴定的语言模型已被证明会遭受堕落的有毒内容和社会偏见行为的损害，从而阻碍了他们的安全部署。提出了各种排毒方法来减轻语言模型的毒性；但是，这些方法是在包含与性别，种族或宗教相关的特定社会身份的提示条件下进行排毒语言模型的。在这项研究中，我们提出了增强氧化。一种基于强化学习的方法，用于降低语言模型中的毒性。我们应对语言模型中的安全性挑战，并提出了一种新的奖励模型，该模型能够检测有毒内容并减轻对毒性预测中社会身份的意外偏见。该实验表明，用于语言模型排毒的增强方法化方法优于自动评估指标中现有的排毒方法，这表明我们在语言模型排毒中的方法能力和对生成内容中社会认同的意外偏见的能力较小。

聊天机器人用于许多应用程序中，例如自动化代理，智能家庭助理，在线游戏中的互动角色等。因此，确保他们不会以不希望的方式行事，对用户提供令人反感或有毒的反应。这并不是一项琐碎的任务，因为最先进的聊天机器人模型是在从互联网公开收集的大型公共数据集上培训的。本文提出了对聊天机器人中毒性的首次大规模测量。我们表明，公开可用的聊天机器人很容易在喂养有毒的查询时提供有毒的反应。更令人担忧的是，一些无毒的查询也会触发有毒反应。然后，我们着手设计和实验攻击，即毒性，该攻击依赖于微调的GPT-2来产生无毒的查询，使聊天机器人以有毒的方式做出反应。我们广泛的实验评估表明，我们的攻击对公共聊天机器人模型有效，并且优于先前工作提出的手动制作的恶意查询。我们还评估了针对毒性的三种防御机制，表明它们要么以影响聊天机器人的效用而降低攻击性能，要么仅有效地减轻了一部分攻击。这强调了对计算机安全和在线安全社区进行更多研究的需求，以确保聊天机器人模型不会伤害其用户。总体而言，我们有信心有毒可以用作审计工具，我们的工作将为设计更有效的聊天机器人安全防御措施铺平道路。

对话模型面临一个困难的权衡。对模型进行了大量文本培训，但是他们的响应需要仅限于对话框的所需范围和样式。由于用于实现前者包含与后者不兼容的语言的数据集，因此在较小的策划数据集上对预训练的对话框模型进行了微调。但是，微调过程剥夺了他们产生多种反应的能力，最终将他们降低到乏味的对话伙伴。在本文中，我们调查提示是否可以减轻上述权衡。具体来说，我们试验了在查询上调节提示，而不是训练所有查询的单个提示。通过遵循直觉，即冻结预训练的语言模型将保持其表现力，我们发现与微调相比，提示可以达到更高的BLEU得分，并显着改善了响应的多样性和新颖性。

良好的善解人意对话系统应首先跟踪并理解用户的情绪，然后以适当的情感回复。但是，目前对此任务的方法要么集中于提高对用户情绪的理解或提出更好的反应策略，而且很少有作品同时考虑这两种工作。我们的工作试图填补这一空缺。受到任务导向对话系统的启发，我们提出了一种具有情感感知对话管理的新颖善解人意的响应生成模型。情绪感知对话管理包含两个部分：（1）情绪状态跟踪保持当前用户的情绪状态，（2）善解人意的对话策略选择预测目标情绪和用户的意图，基于情绪状态跟踪的结果。然后，预测信息用于指导响应的产生。实验结果表明，与自动评估和人类评估下的几个基准相比，动态管理不同的信息可以帮助模型产生更多的移情反应。

We present SODA: the first publicly available, million-scale high-quality social dialogue dataset. Using SODA, we train COSMO: a generalizable conversation agent outperforming previous best-performing agents on both in- and out-of-domain datasets. In contrast to most existing crowdsourced, small-scale dialogue corpora, we distill 1.5M socially-grounded dialogues from a pre-trained language model (InstructGPT; Ouyang et al., 2022). Dialogues are distilled by contextualizing social commonsense knowledge from a knowledge graph (Atomic10x; West et al., 2022). Human evaluation shows that dialogues in SODA are more consistent, specific, and (surprisingly) natural than prior human-authored datasets - e.g., DailyDialog (Li et al., 2017), BlendedSkillTalk (Smith et al., 2020). In addition, extensive evaluations show that COSMO is significantly more natural and consistent on unseen datasets than best-performing dialogue models - e.g., GODEL (Peng et al., 2022), BlenderBot (Roller et al., 2021), DialoGPT (Zhang et al., 2020). Furthermore, it is sometimes even preferred to the original human-written gold responses. We make our data, models, and code public.

深度神经语言模型的最新进展与大规模数据集的能力相结合，加速了自然语言生成系统的发展，这些系统在多种任务和应用程序上下文中产生流利和连贯的文本（在各种成功程度上）。但是，为所需的用户控制这些模型的输出仍然是一个开放的挑战。这不仅对于自定义生成语言的内容和样式至关重要，而且对于他们在现实世界中的安全可靠部署至关重要。我们提出了一项关于受约束神经语言生成的新兴主题的广泛调查，在该主题中，我们通过区分条件和约束（后者是在输出文本上而不是输入的可检验条件），正式定义和分类自然语言生成问题，目前是可检验的）约束文本生成任务，并查看受限文本生成的现有方法和评估指标。我们的目的是强调这个新兴领域的最新进展和趋势，以告知最有希望的方向和局限性，以推动受约束神经语言生成研究的最新作品。

在本文中，我们使用大规模播放脚本数据集来提出从对话中提出戏剧发电的新颖任务。使用超过一百万行的对话和提示，我们将提示生成问题作为受控文本生成任务方法，并展示如何使用如何使用对话/提示鉴别器的语言模型来增强对话的影响。此外，我们还探讨了主题关键字和情绪的使用，以获得受控文本生成。广泛的定量和定性实验表明，语言模型可以成功地用于在高度专业化的域中生成合理的和属性控制的文本，例如播放脚本。配套材料可在：https://catlab-team.github.io/cuegen。

Large pretrained language models can easily produce toxic or biased content, which is prohibitive for practical use. In order to detect such toxic generations, existing methods rely on templates, real-world data extraction, crowdsourcing workers, or automatic generation to construct adversarial contexts that are likely to induce toxic generations. However, what type of context is more likely to induce unsafe responses is still under-explored. In this paper, we identify that context toxicity and context category (e.g., \textit{profanity}, \textit{insult}, \textit{drugs}, etc.) are two important factors to cause safety issues in response generation. Hence, we propose a method called \emph{reverse generation} to construct adversarial contexts conditioned on a given response, with the flexibility to control category, toxicity level, and inductivity of the generated contexts. Via reverse generation, we augment the existing BAD dataset and construct a new dataset BAD+ which contains more than 120K diverse and highly inductive contexts in 12 categories. We test three popular pretrained dialogue models (Blender, DialoGPT, and Plato2) and find that BAD+ can largely expose their safety problems. Furthermore, we show that BAD+ can greatly enhance the safety of generation and reveal the key factors of safety improvement. Our code and dataset is available at \url{https://github.com/thu-coai/Reverse_Generation}.

Natural Language Generation (NLG) represents a large collection of tasks in the field of NLP. While many of these tasks have been tackled well by the cross-entropy (CE) loss, the task of dialog generation poses a few unique challenges for this loss function. First, CE loss assumes that for any given input, the only possible output is the one available as the ground truth in the training dataset. In general, this is not true for any task, as there can be multiple semantically equivalent sentences, each with a different surface form. This problem gets exaggerated further for the dialog generation task, as there can be multiple valid responses (for a given context) that not only have different surface forms but are also not semantically equivalent. Second, CE loss does not take the context into consideration while processing the response and, hence, it treats all ground truths with equal importance irrespective of the context. But, we may want our final agent to avoid certain classes of responses (e.g. bland, non-informative or biased responses) and give relatively higher weightage for more context-specific responses. To circumvent these shortcomings of the CE loss, in this paper, we propose a novel loss function, CORAL, that directly optimizes recently proposed estimates of human preference for generated responses. Using CORAL, we can train dialog generation models without assuming non-existence of response other than the ground-truth. Also, the CORAL loss is computed based on both the context and the response. Extensive comparisons on two benchmark datasets show that the proposed methods outperform strong state-of-the-art baseline models of different sizes.

近年来，文本的风格特性吸引了计算语言学研究人员。具体来说，研究人员研究了文本样式转移（TST）任务，该任务旨在在保留其样式独立内容的同时改变文本的风格属性。在过去的几年中，已经开发了许多新颖的TST算法，而该行业利用这些算法来实现令人兴奋的TST应用程序。由于这种共生，TST研究领域迅速发展。本文旨在对有关文本样式转移的最新研究工作进行全面审查。更具体地说，我们创建了一种分类法来组织TST模型，并提供有关最新技术状况的全面摘要。我们回顾了针对TST任务的现有评估方法，并进行了大规模的可重复性研究，我们在两个公开可用的数据集上实验基准了19个最先进的TST TST算法。最后，我们扩展了当前趋势，并就TST领域的新开发发展提供了新的观点。

The goal of building dialogue agents that can converse with humans naturally has been a long-standing dream of researchers since the early days of artificial intelligence. The well-known Turing Test proposed to judge the ultimate validity of an artificial intelligence agent on the indistinguishability of its dialogues from humans'. It should come as no surprise that human-level dialogue systems are very challenging to build. But, while early effort on rule-based systems found limited success, the emergence of deep learning enabled great advance on this topic. In this thesis, we focus on methods that address the numerous issues that have been imposing the gap between artificial conversational agents and human-level interlocutors. These methods were proposed and experimented with in ways that were inspired by general state-of-the-art AI methodologies. But they also targeted the characteristics that dialogue systems possess.

结合PersonAs信息允许在对话响应生成中多样化和接触响应。不幸的是，事先作品主要专注于自我的人物，并忽视了合作伙伴角色的价值。此外，在实际应用中，实际伙伴角色的可用性通常不是这种情况。本文试图通过提供一种新颖的框架来解决这些问题，这些框架利用自动合作伙伴角色生成来增强成功的对话一代。我们将强化学习纳入了一个专门设计的批评网络，以获得奖励判断。自动和人类评估的实验结果表明a）我们的框架能够产生相关，信息丰富的合作伙伴角色，甚至与地面真理合作伙伴角色相比。 b）生成的合作伙伴角色增强了后续的响应生成，从而超越了当在推理阶段缺少合作伙伴角色时超越了我们的基线和比较模型。 c）我们的框架在推理期间产生的响应比我们的基线在地面真理合作伙伴角色上的基线更具信息丰富和参与。 d）我们专门设计的批评批评网络有效地加强了我们的框架。最后，我们的框架提供了更好的解释性，并降低了对伙伴角色的外部数据库的需求。

Empathy is a vital factor that contributes to mutual understanding, and joint problem-solving. In recent years, a growing number of studies have recognized the benefits of empathy and started to incorporate empathy in conversational systems. We refer to this topic as empathetic conversational systems. To identify the critical gaps and future opportunities in this topic, this paper examines this rapidly growing field using five review dimensions: (i) conceptual empathy models and frameworks, (ii) adopted empathy-related concepts, (iii) datasets and algorithmic techniques developed, (iv) evaluation strategies, and (v) state-of-the-art approaches. The findings show that most studies have centered on the use of the EMPATHETICDIALOGUES dataset, and the text-based modality dominates research in this field. Studies mainly focused on extracting features from the messages of the users and the conversational systems, with minimal emphasis on user modeling and profiling. Notably, studies that have incorporated emotion causes, external knowledge, and affect matching in the response generation models, have obtained significantly better results. For implementation in diverse real-world settings, we recommend that future studies should address key gaps in areas of detecting and authenticating emotions at the entity level, handling multimodal inputs, displaying more nuanced empathetic behaviors, and encompassing additional dialogue system features.