We explore using the 6B parameter GPT-J language model to create a plan for a simulated robot to achieve 30 classes of goals in ScienceWorld, a text game simulator for elementary science experiments and for which previously published empirical work has shown large language models (LLM)s to be a poor fit (Wang et al., 2022). Using the Markov assumption, the LLM outperforms the state-of-the-art based on reinforcement learning by a factor of 1.4. When we fill the LLM’s input buffer with as many prior steps as will fit, improvement rises to 3.3x. Even when training on only 6.5% of the training data, we observe a 2.3x improvement over the state-of-the-art. Our experiments show that performance varies widely across the 30 classes of actions, indicating that averaging over tasks can hide significant performance issues.

Multimodal counterfactual reasoning is a vital yet challenging ability for AI systems. It involves predicting the outcomes of hypothetical circumstances based on vision and language inputs, which enables AI models to learn from failures and explore hypothetical scenarios. Despite its importance, there are only a few datasets targeting the counterfactual reasoning abilities of multimodal models. Among them, they only cover reasoning over synthetic environments or specific types of events (e.g. traffic collisions), making them hard to reliably benchmark the model generalization ability in diverse real-world scenarios and reasoning dimensions. To overcome these limitations, we develop a video question answering dataset, ACQUIRED: it consists of 3.9K annotated videos, encompassing a wide range of event types and incorporating both first and third-person viewpoints, which ensures a focus on real-world diversity. In addition, each video is annotated with questions that span three distinct dimensions of reasoning, including physical, social, and temporal, which can comprehensively evaluate the model counterfactual abilities along multiple aspects. We benchmark our dataset against several state-of-the-art language-only and multimodal models and experimental results demonstrate a significant performance gap (>13%) between models and humans. The findings suggest that multimodal counterfactual reasoning remains an open challenge and ACQUIRED is a comprehensive and reliable benchmark for inspiring future research in this direction.

The ability to sequence unordered events is evidence of comprehension and reasoning about real world tasks/procedures. It is essential for applications such as task planning and multi-source instruction summarization. It often requires thorough understanding of temporal common sense and multimodal information, since these procedures are often conveyed by a combination of texts and images. While humans are capable of reasoning about and sequencing unordered procedural instructions, the extent to which the current machine learning methods possess such capability is still an open question. In this work, we benchmark models’ capability of reasoning over and sequencing unordered multimodal instructions by curating datasets from online instructional manuals and collecting comprehensive human annotations. We find current state-of-the-art models not only perform significantly worse than humans but also seem incapable of efficiently utilizing multimodal information. To improve machines’ performance on multimodal event sequencing, we propose sequence-aware pretraining techniques exploiting the sequential alignment properties of both texts and images, resulting in > 5% improvements on perfect match ratio.

With the recent advances of open-domain story generation, the lack of reliable automatic evaluation metrics becomes an increasingly imperative issue that hinders the fast development of story generation. According to conducted researches in this regard, learnable evaluation metrics have promised more accurate assessments by having higher correlations with human judgments. A critical bottleneck of obtaining a reliable learnable evaluation metric is the lack of high-quality training data for classifiers to efficiently distinguish plausible and implausible machine-generated stories. Previous works relied on heuristically manipulated plausible examples to mimic possible system drawbacks such as repetition, contradiction, or irrelevant content in the text level, which can be unnatural and oversimplify the characteristics of implausible machine-generated stories. We propose to tackle these issues by generating a more comprehensive set of implausible stories using plots, which are structured representations of controllable factors used to generate stories. Since these plots are compact and structured, it is easier to manipulate them to generate text with targeted undesirable properties, while at the same time maintain the grammatical correctness and naturalness of the generated sentences. To improve the quality of generated implausible stories, we further apply the adversarial filtering procedure presented by (CITATION) to select a more nuanced set of implausible texts. Experiments show that the evaluation metrics trained on our generated data result in more reliable automatic assessments that correlate remarkably better with human judgments compared to the baselines.

We describe Machine-Aided Script Curator (MASC), a system for human-machine collaborative script authoring. Scripts produced with MASC include (1) English descriptions of sub-events that comprise a larger, complex event; (2) event types for each of those events; (3) a record of entities expected to participate in multiple sub-events; and (4) temporal sequencing between the sub-events. MASC automates portions of the script creation process with suggestions for event types, links to Wikidata, and sub-events that may have been forgotten. We illustrate how these automations are useful to the script writer with a few case-study scripts.

Our goal is to deliver a new task and leaderboard to stimulate research on question answering and pre-trained language models (PTLMs) to understand a significant instructional document, e.g., an introductory college textbook or a manual. PTLMs have shown great success in many question-answering tasks, given significant supervised training, but much less so in zero-shot settings. We propose a new task that includes two college-level introductory texts in the social sciences (American Government 2e) and humanities (U.S. History), hundreds of true/false statements based on review questions written by the textbook authors, validation/development tests based on the first eight chapters of the textbooks, blind tests based on the remaining textbook chapters, and baseline results given state-of-the-art PTLMs. Since the questions are balanced, random performance should be ~50%. T5, fine-tuned with BoolQ achieves the same performance, suggesting that the textbook’s content is not pre-represented in the PTLM. Taking the exam closed book, but having read the textbook (i.e., adding the textbook to T5’s pre-training), yields at best minor improvement (56%), suggesting that the PTLM may not have “understood” the textbook (or perhaps misunderstood the questions). Performance is better (~60%) when the exam is taken open-book (i.e., allowing the machine to automatically retrieve a paragraph and use it to answer the question).

We consider problems of making sequences of decisions to accomplish tasks, interacting via the medium of language. These problems are often tackled with reinforcement learning approaches. We find that these models do not generalize well when applied to novel task domains. However, the large amount of computation necessary to adequately train and explore the search space of sequential decision making, under a reinforcement learning paradigm, precludes the inclusion of large contextualized language models, which might otherwise enable the desired generalization ability. We introduce a teacher-student imitation learning methodology and a means of converting a reinforcement learning model into a natural language understanding model. Together, these methodologies enable the introduction of contextualized language models into the sequential decision making problem space. We show that models can learn faster and generalize more, leveraging both the imitation learning and the reformulation. Our models exceed teacher performance on various held-out decision problems, by up to 7% on in-domain problems and 24% on out-of-domain problems.

Long-form narrative text generated from large language models manages a fluent impersonation of human writing, but only at the local sentence level, and lacks structure or global cohesion. We posit that many of the problems of story generation can be addressed via high-quality content planning, and present a system that focuses on how to learn good plot structures to guide story generation. We utilize a plot-generation language model along with an ensemble of rescoring models that each implement an aspect of good story-writing as detailed in Aristotle’s Poetics. We find that stories written with our more principled plot-structure are both more relevant to a given prompt and higher quality than baselines that do not content plan, or that plan in an unprincipled way.

We propose a novel deep structured learning framework for event temporal relation extraction. The model consists of 1) a recurrent neural network (RNN) to learn scoring functions for pair-wise relations, and 2) a structured support vector machine (SSVM) to make joint predictions. The neural network automatically learns representations that account for long-term contexts to provide robust features for the structured model, while the SSVM incorporates domain knowledge such as transitive closure of temporal relations as constraints to make better globally consistent decisions. By jointly training the two components, our model combines the benefits of both data-driven learning and knowledge exploitation. Experimental results on three high-quality event temporal relation datasets (TCR, MATRES, and TB-Dense) demonstrate that incorporated with pre-trained contextualized embeddings, the proposed model achieves significantly better performances than the state-of-the-art methods on all three datasets. We also provide thorough ablation studies to investigate our model.

Though information extraction (IE) research has more than a 25-year history, F1 scores remain low. Thus, one could question continued investment in IE research. In this article, we present three applications where information extraction of entities, relations, and/or events has been used, and note the common features that seem to have led to success. We also identify key research challenges whose solution seems essential for broader successes. Because a few practical deployments already exist and because breakthroughs on particular challenges would greatly broaden the technology’s deployment, further R and D investments are justified.

Typically, relation extraction models are trained to extract instances of a relation ontology using only training data from a single language. However, the concepts represented by the relation ontology (e.g. ResidesIn, EmployeeOf) are language independent. The numbers of annotated examples available for a given ontology vary between languages. For example, there are far fewer annotated examples in Spanish and Japanese than English and Chinese. Furthermore, using only language-specific training data results in the need to manually annotate equivalently large amounts of training for each new language a system encounters. We propose a deep neural network to learn transferable, discriminative bilingual representation. Experiments on the ACE 2005 multilingual training corpus demonstrate that the joint training process results in significant improvement in relation classification performance over the monolingual counterparts. The learnt representation is discriminative and transferable between languages. When using 10% (25K English words, or 30K Chinese characters) of the training data, our approach results in doubling F1 compared to a monolingual baseline. We achieve comparable performance to the monolingual system trained with 250K English words (or 300K Chinese characters) With 50% of training data.