Frozen models trained to mimic static datasets can never improve their performance. Models that can employ internet-retrieval for up-to-date information and obtain feedback from humans during deployment provide the promise of both adapting to new information, and improving their performance. In this work we study how to improve internet-driven conversational skills in such a learning framework. We collect deployment data, which we make publicly available, of human interactions, and collect various types of human feedback – including binary quality measurements, free-form text feedback, and fine-grained reasons for failure. We then study various algorithms for improving from such feedback, including standard supervised learning, rejection sampling, model-guiding and reward-based learning, in order to make recommendations on which type of feed- back and algorithms work best. We find the recently introduced DIRECTOR model (Arora et al., 2022) shows significant improvements over other existing approaches.

Many cognitive approaches to well-being, such as recognizing and reframing unhelpful thoughts, have received considerable empirical support over the past decades, yet still lack truly widespread adoption in self-help format. A barrier to that adoption is a lack of adequately specific and diverse dedicated practice material. This work examines whether current language models can be leveraged to both produce a virtually unlimited quantity of practice material illustrating standard unhelpful thought patterns matching specific given contexts, and generate suitable positive reframing proposals. We propose PATTERNREFRAME, a novel dataset of about 10k examples of thoughts containing unhelpful thought patterns conditioned on a given persona, accompanied by about 27k positive reframes. By using this dataset to train and/or evaluate current models, we show that existing models can already be powerful tools to help generate an abundance of tailored practice material and hypotheses, with no or minimal additional model training required.

The social impact of natural language processing and its applications has received increasing attention. In this position paper, we focus on the problem of safety for end-to-end conversational AI. We survey the problem landscape therein, introducing a taxonomy of three observed phenomena: the Instigator, Yea-Sayer, and Impostor effects. We then empirically assess the extent to which current tools can measure these effects and current systems display them. We release these tools as part of a “first aid kit” (SafetyKit) to quickly assess apparent safety concerns. Our results show that, while current tools are able to provide an estimate of the relative safety of systems in various settings, they still have several shortcomings. We suggest several future directions and discuss ethical considerations.

Current open-domain conversational models can easily be made to talk in inadequate ways. Online learning from conversational feedback given by the conversation partner is a promising avenue for a model to improve and adapt, so as to generate fewer of these safety failures. However, current state-of-the-art models tend to react to feedback with defensive or oblivious responses. This makes for an unpleasant experience and may discourage conversation partners from giving feedback in the future. This work proposes SaFeRDialogues, a task and dataset of graceful responses to conversational feedback about safety failures. We collect a dataset of 8k dialogues demonstrating safety failures, feedback signaling them, and a response acknowledging the feedback. We show how fine-tuning on this dataset results in conversations that human raters deem considerably more likely to lead to a civil conversation, without sacrificing engagingness or general conversational ability.

At the heart of improving conversational AI is the open problem of how to evaluate conversations. Issues with automatic metrics are well known (Liu et al., 2016), with human evaluations still considered the gold standard. Unfortunately, how to perform human evaluations is also an open problem: differing data collection methods have varying levels of human agreement and statistical sensitivity, resulting in differing amounts of human annotation hours and labor costs. In this work we compare five different crowdworker-based human evaluation methods and find that different methods are best depending on the types of models compared, with no clear winner across the board. While this highlights the open problems in the area, our analysis leads to advice of when to use which one, and possible future directions.

Over the last several years, end-to-end neural conversational agents have vastly improved their ability to carry unrestricted, open-domain conversations with humans. However, these models are often trained on large datasets from the Internet and, as a result, may learn undesirable behaviours from this data, such as toxic or otherwise harmful language. Thus, researchers must wrestle with how and when to release these models. In this paper, we survey recent and related work to highlight tensions between values, potential positive impact, and potential harms. We also provide a framework to support practitioners in deciding whether and how to release these models, following the tenets of value-sensitive design.

While improving neural dialogue agents’ factual accuracy is the object of much research, another important aspect of communication, less studied in the setting of neural dialogue, is transparency about ignorance. In this work, we analyze to what extent state-of-the-art chit-chat models are linguistically calibrated in the sense that their verbalized expression of doubt (or confidence) matches the likelihood that the model’s responses are factually incorrect (or correct). We find that these models are poorly calibrated, yet we show that likelihood of correctness can accurately be predicted. By incorporating such metacognitive features into the training of a controllable generation model, we obtain a dialogue agent with greatly improved linguistic calibration.

Building open-domain chatbots is a challenging area for machine learning research. While prior work has shown that scaling neural models in the number of parameters and the size of the data they are trained on gives improved results, we highlight other ingredients. Good conversation requires blended skills: providing engaging talking points, and displaying knowledge, empathy and personality appropriately, while maintaining a consistent persona. We show that large scale models can learn these skills when given appropriate training data and choice of generation strategy. We build variants of these recipes with 90M, 2.7B and 9.4B parameter models, and make our models and code publicly available. Human evaluations show our best models outperform existing approaches in multi-turn dialogue on engagingness and humanness measurements. We then discuss the limitations of this work by analyzing failure cases of our models.

Conversational agents trained on large unlabeled corpora of human interactions will learn patterns and mimic behaviors therein, which include offensive or otherwise toxic behavior. We introduce a new human-and-model-in-the-loop framework for evaluating the toxicity of such models, and compare a variety of existing methods in both the cases of non-adversarial and adversarial users that expose their weaknesses. We then go on to propose two novel methods for safe conversational agents, by either training on data from our new human-and-model-in-the-loop framework in a two-stage system, or ”baking-in” safety to the generative model itself. We find our new techniques are (i) safer than existing models; while (ii) maintaining usability metrics such as engagingness relative to state-of-the-art chatbots. In contrast, we expose serious safety issues in existing standard systems like GPT2, DialoGPT, and BlenderBot.

Being engaging, knowledgeable, and empathetic are all desirable general qualities in a conversational agent. Previous work has introduced tasks and datasets that aim to help agents to learn those qualities in isolation and gauge how well they can express them. But rather than being specialized in one single quality, a good open-domain conversational agent should be able to seamlessly blend them all into one cohesive conversational flow. In this work, we investigate several ways to combine models trained towards isolated capabilities, ranging from simple model aggregation schemes that require minimal additional training, to various forms of multi-task training that encompass several skills at all training stages. We further propose a new dataset, BlendedSkillTalk, to analyze how these capabilities would mesh together in a natural conversation, and compare the performance of different architectures and training schemes. Our experiments show that multi-tasking over several tasks that focus on particular capabilities results in better blended conversation performance compared to models trained on a single skill, and that both unified or two-stage approaches perform well if they are constructed to avoid unwanted bias in skill selection or are fine-tuned on our new task.

We introduce dodecaDialogue: a set of 12 tasks that measures if a conversational agent can communicate engagingly with personality and empathy, ask questions, answer questions by utilizing knowledge resources, discuss topics and situations, and perceive and converse about images. By multi-tasking on such a broad large-scale set of data, we hope to both move towards and measure progress in producing a single unified agent that can perceive, reason and converse with humans in an open-domain setting. We show that such multi-tasking improves over a BERT pre-trained baseline, largely due to multi-tasking with very large dialogue datasets in a similar domain, and that the multi-tasking in general provides gains to both text and image-based tasks using several metrics in both the fine-tune and task transfer settings. We obtain state-of-the-art results on many of the tasks, providing a strong baseline for this challenge.

Generative dialogue models currently suffer from a number of problems which standard maximum likelihood training does not address. They tend to produce generations that (i) rely too much on copying from the context, (ii) contain repetitions within utterances, (iii) overuse frequent words, and (iv) at a deeper level, contain logical flaws. In this work we show how all of these problems can be addressed by extending the recently introduced unlikelihood loss (Welleck et al., 2019) to these cases. We show that appropriate loss functions which regularize generated outputs to match human distributions are effective for the first three issues. For the last important general issue, we show applying unlikelihood to collected data of what a model should not do is effective for improving logical consistency, potentially paving the way to generative models with greater reasoning ability. We demonstrate the efficacy of our approach across several dialogue tasks.

Traditional recommendation systems produce static rather than interactive recommendations invariant to a user’s specific requests, clarifications, or current mood, and can suffer from the cold-start problem if their tastes are unknown. These issues can be alleviated by treating recommendation as an interactive dialogue task instead, where an expert recommender can sequentially ask about someone’s preferences, react to their requests, and recommend more appropriate items. In this work, we collect a goal-driven recommendation dialogue dataset (GoRecDial), which consists of 9,125 dialogue games and 81,260 conversation turns between pairs of human workers recommending movies to each other. The task is specifically designed as a cooperative game between two players working towards a quantifiable common goal. We leverage the dataset to develop an end-to-end dialogue system that can simultaneously converse and recommend. Models are first trained to imitate the behavior of human players without considering the task goal itself (supervised training). We then finetune our models on simulated bot-bot conversations between two paired pre-trained models (bot-play), in order to achieve the dialogue goal. Our experiments show that models finetuned with bot-play learn improved dialogue strategies, reach the dialogue goal more often when paired with a human, and are rated as more consistent by humans compared to models trained without bot-play. The dataset and code are publicly available through the ParlAI framework.

Theory of mind, i.e., the ability to reason about intents and beliefs of agents is an important task in artificial intelligence and central to resolving ambiguous references in natural language dialogue. In this work, we revisit the evaluation of theory of mind through question answering. We show that current evaluation methods are flawed and that existing benchmark tasks can be solved without theory of mind due to dataset biases. Based on prior work, we propose an improved evaluation protocol and dataset in which we explicitly control for data regularities via a careful examination of the answer space. We show that state-of-the-art methods which are successful on existing benchmarks fail to solve theory-of-mind tasks in our proposed approach.

One challenge for dialogue agents is recognizing feelings in the conversation partner and replying accordingly, a key communicative skill. While it is straightforward for humans to recognize and acknowledge others’ feelings in a conversation, this is a significant challenge for AI systems due to the paucity of suitable publicly-available datasets for training and evaluation. This work proposes a new benchmark for empathetic dialogue generation and EmpatheticDialogues, a novel dataset of 25k conversations grounded in emotional situations. Our experiments indicate that dialogue models that use our dataset are perceived to be more empathetic by human evaluators, compared to models merely trained on large-scale Internet conversation data. We also present empirical comparisons of dialogue model adaptations for empathetic responding, leveraging existing models or datasets without requiring lengthy re-training of the full model.