Action knowledge involves the understanding of textual, visual, and temporal aspects of actions. We introduce the Action Dynamics Benchmark (ActionBench) containing two carefully designed probing tasks: Action Antonym and Video Reversal, which targets multimodal alignment capabilities and temporal understanding skills of the model, respectively. Despite recent video-language models' (VidLM) impressive performance on various benchmark tasks, our diagnostic tasks reveal their surprising deficiency (near-random performance) in action knowledge, suggesting that current models rely on object recognition abilities as a shortcut for action understanding. To remedy this, we propose a novel framework, Paxion, along with a new Discriminative Video Dynamics Modeling (DVDM) objective. The Paxion framework utilizes a Knowledge Patcher network to encode new action knowledge and a Knowledge Fuser component to integrate the Patcher into frozen VidLMs without compromising their existing capabilities. Due to limitations of the widely-used Video-Text Contrastive (VTC) loss for learning action knowledge, we introduce the DVDM objective to train the Knowledge Patcher. DVDM forces the model to encode the correlation between the action text and the correct ordering of video frames. Our extensive analyses show that Paxion and DVDM together effectively fill the gap in action knowledge understanding (~50% to 80%), while maintaining or improving performance on a wide spectrum of both object- and action-centric downstream tasks.
Neural information retrieval often adopts a retrieve-and-rerank framework: a bi-encoder network first retrieves K (e.g., 100) candidates that are then re-ranked using a more powerful cross-encoder model to rank the better candidates higher. The re-ranker generally produces better candidate scores than the retriever, but is limited to seeing only the top K retrieved candidates, thus providing no improvements in retrieval performance as measured by Recall@K. In this work, we leverage the re-ranker to also improve retrieval by providing inference-time relevance feedback to the retriever. Concretely, we update the retriever's query representation for a test instance using a lightweight inference-time distillation of the re-ranker's prediction for that instance. The distillation loss is designed to bring the retriever's candidate scores closer to those of the re-ranker. A second retrieval step is then performed with the updated query vector. We empirically show that our approach, which can serve arbitrary retrieve-and-rerank pipelines, significantly improves retrieval recall in multiple domains, languages, and modalities.
Large language Models (LLMs) have achieved promising performance on arithmetic reasoning tasks by incorporating step-by-step chain-of-thought (CoT) prompting. However, LLMs face challenges in maintaining factual consistency during reasoning, exhibiting tendencies to condition overlooking, question misinterpretation, and condition hallucination over given problems. Existing methods use coarse-grained feedback (e.g., whether the answer is correct) to improve factual consistency. In this work, we propose RCoT (Reversing Chain-of-Thought), a novel method to improve LLMs' reasoning abilities by automatically detecting and rectifying factual inconsistency in LLMs' generated solutions. To detect factual inconsistency, RCoT first asks LLMs to reconstruct the problem based on generated solutions. Then fine-grained comparisons between the original problem and the reconstructed problem expose the factual inconsistency in the original solutions. To rectify the solution, RCoT formulates detected factual inconsistency into fine-grained feedback to guide LLMs in revising solutions. Experimental results demonstrate consistent improvements of RCoT over standard CoT across seven arithmetic datasets. Moreover, we find that manually written fine-grained feedback can dramatically improve LLMs' reasoning abilities (e.g., ChatGPT reaches 94.6% accuracy on GSM8K), encouraging the community to further explore the fine-grained feedback generation methods.
Finetuning pre-trained language models (LMs) enhances the models' capabilities. Prior techniques fine-tune a pre-trained LM on input-output pairs (e.g., instruction fine-tuning), or with numerical rewards that gauge the quality of its outputs (e.g., reinforcement learning from human feedback). We explore LMs' potential to learn from textual interactions (LeTI) that not only check their correctness with binary labels, but also pinpoint and explain errors in their outputs through textual feedback. Our investigation focuses on the code generation task, where the model produces code pieces in response to natural language instructions. This setting invites a natural and scalable way to acquire the textual feedback: the error messages and stack traces from code execution using a Python interpreter. LeTI iteratively fine-tunes the model, using the LM objective, on a concatenation of natural language instructions, LM-generated programs, and textual feedback, which is only provided when the generated program fails to solve the task. Prepended to this fine-tuning text, a binary reward token is used to differentiate correct and buggy solutions. On MBPP, a code generation dataset, LeTI substantially improves the performance of two base LMs of different scales. LeTI requires no ground-truth outputs for training and even outperforms a fine-tuned baseline that does. LeTI's strong performance generalizes to other datasets. Trained on MBPP, it achieves comparable or better performance than the base LMs on unseen problems in HumanEval. Furthermore, compared to binary feedback, we observe that textual feedback leads to improved generation quality and sample efficiency, achieving the same performance with fewer than half of the gradient steps. LeTI is equally applicable in natural language tasks when they can be formulated as code generation, which we empirically verified on event argument extraction.
Faithfully correcting factual errors is critical for maintaining the integrity of textual knowledge bases and preventing hallucinations in sequence-to-sequence models. Drawing on humans' ability to identify and correct factual errors, we present a zero-shot framework that formulates questions about input claims, looks for correct answers in the given evidence, and assesses the faithfulness of each correction based on its consistency with the evidence. Our zero-shot framework outperforms fully-supervised approaches, as demonstrated by experiments on the FEVER and SciFact datasets, where our outputs are shown to be more faithful. More importantly, the decomposability nature of our framework inherently provides interpretability. Additionally, to reveal the most suitable metrics for evaluating factual error corrections, we analyze the correlation between commonly used metrics with human judgments in terms of three different dimensions regarding intelligibility and faithfulness.
Humans possess an extraordinary ability to create and utilize tools, allowing them to overcome physical limitations and explore new frontiers. With the advent of foundation models, AI systems have the potential to be equally adept in tool use as humans. This paradigm, i.e., tool learning with foundation models, combines the strengths of specialized tools and foundation models to achieve enhanced accuracy, efficiency, and automation in problem-solving. Despite its immense potential, there is still a lack of a comprehensive understanding of key challenges, opportunities, and future endeavors in this field. To this end, we present a systematic investigation of tool learning in this paper. We first introduce the background of tool learning, including its cognitive origins, the paradigm shift of foundation models, and the complementary roles of tools and models. Then we recapitulate existing tool learning research into tool-augmented and tool-oriented learning. We formulate a general tool learning framework: starting from understanding the user instruction, models should learn to decompose a complex task into several subtasks, dynamically adjust their plan through reasoning, and effectively conquer each sub-task by selecting appropriate tools. We also discuss how to train models for improved tool-use capabilities and facilitate the generalization in tool learning. Considering the lack of a systematic tool learning evaluation in prior works, we experiment with 17 representative tools and show the potential of current foundation models in skillfully utilizing tools. Finally, we discuss several open problems that require further investigation for tool learning. Overall, we hope this paper could inspire future research in integrating tools with foundation models.
Emerging events, such as the COVID pandemic and the Ukraine Crisis, require a time-sensitive comprehensive understanding of the situation to allow for appropriate decision-making and effective action response. Automated generation of situation reports can significantly reduce the time, effort, and cost for domain experts when preparing their official human-curated reports. However, AI research toward this goal has been very limited, and no successful trials have yet been conducted to automate such report generation. We propose SmartBook, a novel task formulation targeting situation report generation, which consumes large volumes of news data to produce a structured situation report with multiple hypotheses (claims) summarized and grounded with rich links to factual evidence. We realize SmartBook for the Ukraine-Russia crisis by automatically generating intelligence analysis reports to assist expert analysts. The machine-generated reports are structured in the form of timelines, with each timeline organized by major events (or chapters), corresponding strategic questions (or sections) and their grounded summaries (or section content). Our proposed framework automatically detects real-time event-related strategic questions, which are more directed than manually-crafted analyst questions, which tend to be too complex, hard to parse, vague and high-level. Results from thorough qualitative evaluations show that roughly 82% of the questions in Smartbook have strategic importance, with at least 93% of the sections in the report being tactically useful. Further, experiments show that expert analysts tend to add more information into the SmartBook reports, with only 2.3% of the existing tokens being deleted, meaning SmartBook can serve as a useful foundation for analysts to build upon when creating intelligence reports.
The development of event extraction systems has been hindered by the absence of wide-coverage, large-scale datasets. To make event extraction systems more accessible, we build a general-purpose event detection dataset GLEN, which covers 3,465 different event types, making it over 20x larger in ontology than any current dataset. GLEN is created by utilizing the DWD Overlay, which provides a mapping between Wikidata Qnodes and PropBank rolesets. This enables us to use the abundant existing annotation for PropBank as distant supervision. In addition, we also propose a new multi-stage event detection model specifically designed to handle the large ontology size and partial labels in GLEN. We show that our model exhibits superior performance (~10% F1 gain) compared to both conventional classification baselines and newer definition-based models. Finally, we perform error analysis and show that label noise is still the largest challenge for improving performance.
Schema induction builds a graph representation explaining how events unfold in a scenario. Existing approaches have been based on information retrieval (IR) and information extraction(IE), often with limited human curation. We demonstrate a human-in-the-loop schema induction system powered by GPT-3. We first describe the different modules of our system, including prompting to generate schematic elements, manual edit of those elements, and conversion of those into a schema graph. By qualitatively comparing our system to previous ones, we show that our system not only transfers to new domains more easily than previous approaches, but also reduces efforts of human curation thanks to our interactive interface.
Entity coreference resolution is an important research problem with many applications, including information extraction and question answering. Coreference resolution for English has been studied extensively. However, there is relatively little work for other languages. A problem that frequently occurs when working with a non-English language is the scarcity of annotated training data. To overcome this challenge, we design a simple but effective ensemble-based framework that combines various transfer learning (TL) techniques. We first train several models using different TL methods. Then, during inference, we compute the unweighted average scores of the models' predictions to extract the final set of predicted clusters. Furthermore, we also propose a low-cost TL method that bootstraps coreference resolution models by utilizing Wikipedia anchor texts. Leveraging the idea that the coreferential links naturally exist between anchor texts pointing to the same article, our method builds a sizeable distantly-supervised dataset for the target language that consists of tens of thousands of documents. We can pre-train a model on the pseudo-labeled dataset before finetuning it on the final target dataset. Experimental results on two benchmark datasets, OntoNotes and SemEval, confirm the effectiveness of our methods. Our best ensembles consistently outperform the baseline approach of simple training by up to 7.68% in the F1 score. These ensembles also achieve new state-of-the-art results for three languages: Arabic, Dutch, and Spanish.