Large, transformer-based pretrained language models like BERT, GPT, and T5 have demonstrated a deep understanding of contextual semantics and language syntax. Their success has enabled significant advances in conversational AI, including the development of open-dialogue systems capable of coherent, salient conversations which can answer questions, chat casually, and complete tasks. However, state-of-the-art models still struggle with tasks that involve higher levels of reasoning - including commonsense reasoning that humans find trivial. This paper presents a survey of recent conversational AI research focused on commonsense reasoning. The paper lists relevant training datasets and describes the primary approaches to include commonsense in conversational AI. The paper also discusses benchmarks used for evaluating commonsense in conversational AI problems. Finally, the paper presents preliminary observations of the limited commonsense capabilities of two state-of-the-art open dialogue models, BlenderBot3 and LaMDA, and its negative effect on natural interactions. These observations further motivate research on commonsense reasoning in conversational AI.

A common limitation of diagnostic tests for detecting social biases in NLP models is that they may only detect stereotypic associations that are pre-specified by the designer of the test. Since enumerating all possible problematic associations is infeasible, it is likely these tests fail to detect biases that are present in a model but not pre-specified by the designer. To address this limitation, we propose SODAPOP (SOcial bias Discovery from Answers about PeOPle) in social commonsense question-answering. Our pipeline generates modified instances from the Social IQa dataset (Sap et al., 2019) by (1) substituting names associated with different demographic groups, and (2) generating many distractor answers from a masked language model. By using a social commonsense model to score the generated distractors, we are able to uncover the model's stereotypic associations between demographic groups and an open set of words. We also test SODAPOP on debiased models and show the limitations of multiple state-of-the-art debiasing algorithms.

In this work, we explore techniques for augmenting interactive agents with information from symbolic modules, much like humans use tools like calculators and GPS systems to assist with arithmetic and navigation. We test our agent's abilities in text games -- challenging benchmarks for evaluating the multi-step reasoning abilities of game agents in grounded, language-based environments. Our experimental study indicates that injecting the actions from these symbolic modules into the action space of a behavior cloned transformer agent increases performance on four text game benchmarks that test arithmetic, navigation, sorting, and common sense reasoning by an average of 22%, allowing an agent to reach the highest possible performance on unseen games. This action injection technique is easily extended to new agents, environments, and symbolic modules.

Contextualized or discourse aware commonsense inference is the task of generating coherent commonsense assertions (i.e., facts) from a given story, and a particular sentence from that story. Some problems with the task are: lack of controllability for topics of the inferred facts; lack of commonsense knowledge during training; and, possibly, hallucinated or false facts. In this work, we utilize a transformer model for this task and develop techniques to address the aforementioned problems in the task. We control the inference by introducing a new technique we call "hinting". Hinting is a kind of language model prompting, that utilizes both hard prompts (specific words) and soft prompts (virtual learnable templates). This serves as a control signal to advise the language model "what to talk about". Next, we establish a methodology for performing joint inference with multiple commonsense knowledge bases. Joint inference of commonsense requires care, because it is imprecise and the level of generality is more flexible. You want to be sure that the results "still make sense" for the context. To this end, we align the textual version of assertions from three knowledge graphs (ConceptNet, ATOMIC2020, and GLUCOSE) with a story and a target sentence. This combination allows us to train a single model to perform joint inference with multiple knowledge graphs. We show experimental results for the three knowledge graphs on joint inference. Our final contribution is exploring a GAN architecture that generates the contextualized commonsense assertions and scores them as to their plausibility through a discriminator. The result is an integrated system for contextual commonsense inference in stories, that can controllably generate plausible commonsense assertions, and takes advantage of joint inference between multiple commonsense knowledge bases.

Improving the emotional awareness of pre-trained language models is an emerging important problem for dialogue generation tasks. Although prior studies have introduced methods to improve empathetic dialogue generation, few have discussed how to incorporate commonsense knowledge into pre-trained language models for controllable dialogue generation. In this study, we propose a novel framework that improves empathetic dialogue generation using pre-trained language models by 1) incorporating commonsense knowledge through prompt verbalization, and 2) controlling dialogue generation using a strategy-driven future discriminator. We conducted experiments to reveal that both the incorporation of social commonsense knowledge and enforcement of control over generation help to improve generation performance. Finally, we discuss the implications of our study for future research.

The paper discusses the capacities and limitations of current artificial intelligence (AI) technology to solve word problems that combine elementary knowledge with commonsense reasoning. No existing AI systems can solve these reliably. We review three approaches that have been developed, using AI natural language technology: outputting the answer directly, outputting a computer program that solves the problem, and outputting a formalized representation that can be input to an automated theorem verifier. We review some benchmarks that have been developed to evaluate these systems and some experimental studies. We discuss the limitations of the existing technology at solving these kinds of problems. We argue that it is not clear whether these kinds of limitations will be important in developing AI technology for pure mathematical research, but that they will be important in applications of mathematics, and may well be important in developing programs capable of reading and understanding mathematical content written by humans.

The Winograd Schema Challenge - a set of twin sentences involving pronoun reference disambiguation that seem to require the use of commonsense knowledge - was proposed by Hector Levesque in 2011. By 2019, a number of AI systems, based on large pre-trained transformer-based language models and fine-tuned on these kinds of problems, achieved better than 90% accuracy. In this paper, we review the history of the Winograd Schema Challenge and discuss the lasting contributions of the flurry of research that has taken place on the WSC in the last decade. We discuss the significance of various datasets developed for WSC, and the research community's deeper understanding of the role of surrogate tasks in assessing the intelligence of an AI system.

Data Augmentation (DA) is frequently used to automatically provide additional training data without extra human annotation. However, data augmentation may introduce noisy data that impairs training. To guarantee the quality of augmented data, existing methods either assume no noise exists in the augmented data and adopt consistency training or use simple heuristics such as training loss and diversity constraints to filter out ``noisy'' data. However, those filtered examples may still contain useful information, and dropping them completely causes loss of supervision signals. In this paper, based on the assumption that the original dataset is cleaner than the augmented data, we propose an on-the-fly denoising technique for data augmentation that learns from soft augmented labels provided by an organic teacher model trained on the cleaner original data. A simple self-regularization module is applied to force the model prediction to be consistent across two distinct dropouts to further prevent overfitting on noisy labels. Our method can be applied to augmentation techniques in general and can consistently improve the performance on both text classification and question-answering tasks.

Commonsense knowledge-graphs (CKGs) are important resources towards building machines that can 'reason' on text or environmental inputs and make inferences beyond perception. While current CKGs encode world knowledge for a large number of concepts and have been effectively utilized for incorporating commonsense in neural models, they primarily encode declarative or single-condition inferential knowledge and assume all conceptual beliefs to have the same likelihood. Further, these CKGs utilize a limited set of relations shared across concepts and lack a coherent knowledge organization structure resulting in redundancies as well as sparsity across the larger knowledge graph. Consequently, today's CKGs, while useful for a first level of reasoning, do not adequately capture deeper human-level commonsense inferences which can be more nuanced and influenced by multiple contextual or situational factors. Accordingly, in this work, we study how commonsense knowledge can be better represented by -- (i) utilizing a probabilistic logic representation scheme to model composite inferential knowledge and represent conceptual beliefs with varying likelihoods and (ii) incorporating a hierarchical conceptual ontology to identify salient concept-relevant relations and organize beliefs at different conceptual levels. Our resulting knowledge representation framework can encode a wider variety of world knowledge and represent beliefs flexibly using grounded concepts as well as free-text phrases. As a result, the framework can be utilized as both a traditional free-text knowledge graph and a grounded logic-based inference system more suitable for neuro-symbolic applications. We describe how we extend the PrimeNet knowledge base with our framework through crowd-sourcing and expert-annotation, and demonstrate its application for more interpretable passage-based semantic parsing and question answering.

Understanding novel situations in the traffic domain requires an intricate combination of domain-specific and causal commonsense knowledge. Prior work has provided sufficient perception-based modalities for traffic monitoring, in this paper, we focus on a complementary research aspect of Intelligent Transportation: traffic understanding. We scope our study to text-based methods and datasets given the abundant commonsense knowledge that can be extracted using language models from large corpus and knowledge graphs. We adopt three knowledge-driven approaches for zero-shot QA over traffic situations, based on prior natural language inference methods, commonsense models with knowledge graph self-supervision, and dense retriever-based models. We constructed two text-based multiple-choice question answering sets: BDD-QA for evaluating causal reasoning in the traffic domain and HDT-QA for measuring the possession of domain knowledge akin to human driving license tests. Among the methods, Unified-QA reaches the best performance on the BDD-QA dataset with the adaptation of multiple formats of question answers. Language models trained with inference information and commonsense knowledge are also good at predicting the cause and effect in the traffic domain but perform badly at answering human-driving QA sets. For such sets, DPR+Unified-QA performs the best due to its efficient knowledge extraction.