Xinliang Frederick Zhang, Anhad Mohananey, Alexandra Chronopoulou, Pinelopi Papalampidi, Somit Gupta, Tsendsuren Munkhdalai, Lu Wang, Shyam Upadhyay. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

While text-to-image (T2I) generative models have become ubiquitous, they do not necessarily generate images that align with a given prompt. While many metrics and benchmarks have been proposed to evaluate T2I models and alignment metrics, the impact of the evaluation components (prompt sets, human annotations, evaluation task) has not been systematically measured. We find that looking at only *one slice of data*, i.e. one set of capabilities or human annotations, is not enough to obtain stable conclusions that generalise to new conditions or slices when evaluating T2I models or alignment metrics. We address this by introducing an evaluation suite of $>$100K annotations across four human annotation templates that comprehensively evaluates models' capabilities across a range of methods for gathering human annotations and comparing models. In particular, we propose (1) a carefully curated set of prompts -- *Gecko2K*; (2) a statistically grounded method of comparing T2I models; and (3) how to systematically evaluate metrics under three *evaluation tasks* -- *model ordering, pair-wise instance scoring, point-wise instance scoring*. Using this evaluation suite, we evaluate a wide range of metrics and find that a metric may do better in one setting but worse in another. As a result, we introduce a new, interpretable auto-eval metric that is consistently better correlated with human ratings than such existing metrics on our evaluation suite--across different human templates and evaluation settings--and on TIFA160.

Understanding long, real-world videos requires modeling of long-range visual dependencies. To this end, we explore video-first architectures, building on the common paradigm of transferring large-scale, image–text models to video via shallow temporal fusion. However, we expose two limitations to the approach: (1) decreased spatial capabilities, likely due to poor video–language alignment in standard video datasets, and (2) higher memory consumption, bottlenecking the number of frames that can be processed. To mitigate the memory bottleneck, we systematically analyze the memory/accuracy trade-off of various efficient methods: factorized attention, parameter-efficient image-to-video adaptation, input masking, and multi-resolution patchification. Surprisingly, simply masking large portions of the video (up to 75%) during contrastive pre-training proves to be one of the most robust ways to scale encoders to videos up to 4.3 minutes at 1 FPS. Our simple approach for training long video-to-text models, which scales to 1B parameters, does not add new architectural complexity and is able to outperform the popular paradigm of using much larger LLMs as an information aggregator over segment-based information on benchmarks with long-range temporal dependencies (YouCook2, EgoSchema).

Most transformer-based video encoders are limited to short temporal contexts due to their quadratic complexity. While various attempts have been made to extend this context, this has often come at the cost of both conceptual and computational complexity. We propose to instead re-purpose existing pre-trained video transformers by simply fine-tuning them to attend to memories derived non-parametrically from past activations. By leveraging redundancy reduction, our memory-consolidated vision transformer (MC-ViT) effortlessly extends its context far into the past and exhibits excellent scaling behavior when learning from longer videos. In doing so, MC-ViT sets a new state-of-the-art in long-context video understanding on EgoSchema, Perception Test, and Diving48, outperforming methods that benefit from orders of magnitude more parameters.

Movie trailers perform multiple functions: they introduce viewers to the story, convey the mood and artistic style of the film, and encourage audiences to see the movie. These diverse functions make trailer creation a challenging endeavor. In this work, we focus on finding trailer moments in a movie, i.e., shots that could be potentially included in a trailer. We decompose this task into two subtasks: narrative structure identification and sentiment prediction. We model movies as graphs, where nodes are shots and edges denote semantic relations between them. We learn these relations using joint contrastive training which distills rich textual information (e.g., characters, actions, situations) from screenplays. An unsupervised algorithm then traverses the graph and selects trailer moments from the movie that human judges prefer to ones selected by competitive supervised approaches. A main advantage of our algorithm is that it uses interpretable criteria, which allows us to deploy it in an interactive tool for trailer creation with a human in the loop. Our tool allows users to select trailer shots in under 30 minutes that are superior to fully automatic methods and comparable to (exclusive) manual selection by experts.

Large pre-trained language models (LMs) have demonstrated impressive capabilities in generating long, fluent text; however, there is little to no analysis on their ability to maintain entity coherence and consistency. In this work, we focus on the end task of narrative generation and systematically analyse the long-range entity coherence and consistency in generated stories. First, we propose a set of automatic metrics for measuring model performance in terms of entity usage. Given these metrics, we quantify the limitations of current LMs. Next, we propose augmenting a pre-trained LM with a dynamic entity memory in an end-to-end manner by using an auxiliary entity-related loss for guiding the reads and writes to the memory. We demonstrate that the dynamic entity memory increases entity coherence according to both automatic and human judgment and helps preserving entity-related information especially in settings with a limited context window. Finally, we also validate that our automatic metrics are correlated with human ratings and serve as a good indicator of the quality of generated stories.

We summarize full-length movies by creating shorter videos containing their most informative scenes. We explore the hypothesis that a summary can be created by assembling scenes which are turning points (TPs), i.e., key events in a movie that describe its storyline. We propose a model that identifies TP scenes by building a sparse movie graph that represents relations between scenes and is constructed using multimodal information. According to human judges, the summaries created by our approach are more informative and complete, and receive higher ratings, than the outputs of sequence-based models and general-purpose summarization algorithms. The induced graphs are interpretable, displaying different topology for different movie genres.

Most general-purpose extractive summarization models are trained on news articles, which are short and present all important information upfront.As a result, such models are biased by position and often perform a smart selection of sentences from the beginning of the document.When summarizing long narratives, which have complex structure and present information piecemeal, simple position heuristics are not sufficient.In this paper, we propose to explicitly incorporate the underlying structure of narratives into general unsupervised and supervised extractive summarization models.We formalize narrative structure in terms of key narrative events (turning points) and treat it as latent in order to summarize screenplays (i.e., extract an optimal sequence of scenes).Experimental results on the CSI corpus of TV screenplays, which we augment with scene-level summarization labels, show that latent turning points correlate with important aspects of a CSI episode and improve summarization performance over general extractive algorithms, leading to more complete and diverse summaries.Victim: Mike Kimble, found in a Body Farm.Died 6 hours ago, unknown cause of death.CSI discover cow tissue in Mike's body.Cross-contamination is suggested.Probable cause of death: Mike's house has been set on fire.CSI finds blood: Mike was murdered, fire was a cover up.First suspects: Mike's fiance, Jane and her ex-husband, Russ.CSI finds photos in Mike's house of Jane's daughter, Jodie, posing naked.Mike is now a suspect of abusing Jodie.Russ allows CSI to examine his gun.CSI discovers that the bullet that killed Mike was made of frozen beef that melt inside him.They also find beef in Russ' gun.Russ confesses that he knew that Mike was abusing Jody, so he confronted and killed him.CSI discovers that the naked photos were taken on a boat, which belongs to Russ.CSI discovers that it was Russ who was abusing his daughter based on fluids found in his sleeping bag and later killed Mike who tried to help Jodie.Russ is given bail, since no jury would convict a protective father.Russ receives a mandatory life sentence.

Pinelopi Papalampidi, Frank Keller, Mirella Lapata. Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP). 2019.