Search for: All records

We evaluate the ability of large language models (LLMs) to provide PropBank semantic role label annotations across different realizations of the same verbs in transitive, intransitive, and middle voice constructions. In order to assess the meta-linguistic capabilities of LLMs as well as their ability to glean such capabilities through in-context learning, we evaluate the models in a zero-shot setting, in a setting where it is given three examples of another verb used in transitive, intransitive, and middle voice constructions, and finally in a setting where it is given the examples as well as the correct sense and roleset information. We find that zero-shot knowledge of PropBank annotation is almost nonexistent. The largest model evaluated, GPT-4, achieves the best performance in the setting where it is given both examples and the correct roleset in the prompt, demonstrating that larger models can ascertain some meta-linguistic capabilities through in-context learning. However, even in this setting, which is simpler than the task of a human in PropBank annotation, the model achieves only 48% accuracy in marking numbered arguments correctly. To ensure transparency and reproducibility, we publicly release our dataset and model responses. 

Despite considerable advancements with deep neural language models (LMs), neural text generation still suffers from degeneration: the generated text is repetitive, generic, selfcontradictory, and often lacks commonsense. Our analyses on sentence-level attention patterns in LMs reveal that neural degeneration may be associated with insufficient learning of task-specific characteristics by the attention mechanism. This finding motivates onthe-fly attention modulation1– a simple but effective method that enables the injection of priors into attention computation during inference. Automatic and human evaluation results on three text generation benchmarks demonstrate that attention modulation helps LMs generate text with enhanced fluency, creativity, and commonsense reasoning, in addition to significantly reduce sentence-level repetition. 

Multimodal disinformation, from `deepfakes' to simple edits that deceive, is an important societal problem. Yet at the same time, the vast majority of media edits are harmless -- such as a filtered vacation photo. The difference between this example, and harmful edits that spread disinformation, is one of intent. Recognizing and describing this intent is a major challenge for today's AI systems. We present the task of Edited Media Understanding, requiring models to answer open-ended questions that capture the intent and implications of an image edit. We introduce a dataset for our task, EMU, with 48k question-answer pairs written in rich natural language. We evaluate a wide variety of vision-and-language models for our task, and introduce a new model PELICAN, which builds upon recent progress in pretrained multimodal representations. Our model obtains promising results on our dataset, with humans rating its answers as accurate 40.35% of the time. At the same time, there is still much work to be done -- humans prefer human-annotated captions 93.56% of the time -- and we provide analysis that highlights areas for further progress. 

While many languages use adpositions to encode semantic relationships between content words in a sentence (e.g., agentivity or temporality), the details of how adpositions work vary widely across languages with respect to both form and meaning. In this paper, we empirically adapt the SNACS framework (Schneider et al., 2018) to Korean, a language that is typologically distant from English—the language SNACS was based on. We apply the SNACS framework to annotate the highly popular novella The Little Prince with semantic supersense labels over allKorean postpositions. Thus, we introduce the first broad-coverage corpus annotated with Korean postposition semantics and provide a detailed analysis of the corpus with an apples-to-apples comparison between Korean and English annotations.