The type of content material that customers may wish to create utilizing a generative mannequin corresponding to Flux or Hunyuan Video is probably not all the time be simply obtainable, even when the content material request is pretty generic, and one may guess that the generator may deal with it.
One instance, illustrated in a brand new paper that we’ll check out on this article, notes that the increasingly-eclipsed OpenAI Sora mannequin has some issue rendering an anatomically appropriate firefly, utilizing the immediate ‘A firefly is glowing on a grass’s leaf on a serene summer season evening’:
OpenAI’s Sora has a barely wonky understanding of firefly anatomy. Supply: https://arxiv.org/pdf/2503.01739
Since I not often take analysis claims at face worth, I examined the identical immediate on Sora in the present day and acquired a barely higher end result. Nonetheless, Sora nonetheless didn’t render the glow appropriately – slightly than illuminating the tip of the firefly’s tail, the place bioluminescence happens, it misplaced the glow close to the insect’s toes:
My very own take a look at of the researchers’ immediate in Sora produces a end result that exhibits Sora doesn’t perceive the place a Firefly’s gentle really comes from.
Sarcastically, the Adobe Firefly generative diffusion engine, skilled on the corporate’s copyright-secured inventory pictures and movies, solely managed a 1-in-3 success charge on this regard, once I tried the identical immediate in Photoshop’s generative AI characteristic:
Solely the ultimate of three proposed generations of the researchers’ immediate produces a glow in any respect in Adobe Firefly (March 2025), although at the least the glow is located within the appropriate a part of the insect’s anatomy.
This instance was highlighted by the researchers of the brand new paper as an instance that the distribution, emphasis and protection in coaching units used to tell standard basis fashions could not align with the consumer’s wants, even when the consumer shouldn’t be asking for something notably difficult – a subject that brings up the challenges concerned in adapting hyperscale coaching datasets to their most effective and performative outcomes as generative fashions.
The authors state:
‘[Sora] fails to seize the idea of a glowing firefly whereas efficiently producing grass and a summer season [night]. From the information perspective, we infer that is primarily as a result of [Sora] has not been skilled on firefly-related matters, whereas it has been skilled on grass and evening. Moreover, if [Sora had] seen the video proven in [above image], it’s going to perceive what a glowing firefly ought to seem like.’
They introduce a newly curated dataset and counsel that their methodology may very well be refined in future work to create information collections that higher align with consumer expectations than many current fashions.
Knowledge for the Individuals
Primarily their proposal posits a knowledge curation strategy that falls someplace between the customized information for a model-type corresponding to a LoRA (and this strategy is much too particular for common use); and the broad and comparatively indiscriminate high-volume collections (such because the LAION dataset powering Secure Diffusion) which aren’t particularly aligned with any end-use situation.
The brand new strategy, each as methodology and a novel dataset, is (slightly tortuously) named Customers’ FOcus in text-to-video, or VideoUFO. The VideoUFO dataset contains 1.9 million video clips spanning 1291 user-focused matters. The matters themselves had been elaborately developed from an current video dataset, and parsed by way of numerous language fashions and Pure Language Processing (NLP) methods:
Samples of the distilled matters introduced within the new paper.
The VideoUFO dataset includes a excessive quantity of novel movies trawled from YouTube – ‘novel’ within the sense that the movies in query don’t characteristic in video datasets which might be at present standard within the literature, and due to this fact within the many subsets which have been curated from them (and most of the movies had been in truth uploaded subsequent to the creation of the older datasets thar the paper mentions).
In actual fact, the authors declare that there’s solely 0.29% overlap with current video datasets – a formidable demonstration of novelty.
One motive for this could be that the authors would solely settle for YouTube movies with a Inventive Commons license that will be much less prone to hamstring customers additional down the road: it is doable that this class of movies has been much less prioritized in prior sweeps of YouTube and different high-volume platforms.
Secondly, the movies had been requested on the premise of pre-estimated user-need (see picture above), and never indiscriminately trawled. These two components together may result in such a novel assortment. Moreover, the researchers checked the YouTube IDs of any contributing movies (i.e., movies which will later have been break up up and re-imagined for the VideoUFO assortment) towards these featured in current collections, lending credence to the declare.
Although not all the things within the new paper is kind of as convincing, it is an fascinating learn that emphasizes the extent to which we’re nonetheless slightly on the mercy of uneven distributions in datasets, when it comes to the obstacles the analysis scene is commonly confronted with in dataset curation.
The new work is titled VideoUFO: A Million-Scale Person-Targeted Dataset for Textual content-to-Video Technology, and comes from two researchers, respectively from the College of Know-how Sydney in Australia, and Zhejiang College in China.
Choose examples from the ultimate obtained dataset.
A ‘Private Shopper’ for AI Knowledge
The subject material and ideas featured within the complete sum of web pictures and movies don’t essentially replicate what the typical finish consumer could find yourself asking for from a generative system; even the place content material and demand do are likely to collide (as with porn, which is plentifully available on the web and of great interest to many gen AI customers), this may occasionally not align with the builders’ intent and requirements for a brand new generative system.
Moreover the excessive quantity of NSFW materials uploaded each day, a disproportionate quantity of net-available materials is prone to be from advertisers and people attempting to manipulate SEO. Industrial self-interest of this type makes the distribution of subject material removed from neutral; worse, it’s troublesome to develop AI-based filtering methods that may deal with the issue, since algorithms and fashions developed from significant hyperscale information could in themselves replicate the supply information’s tendencies and priorities.
Due to this fact the authors of the brand new work have approached the issue by reversing the proposition, by way of figuring out what customers are prone to need, and acquiring movies that align with these wants.
On the floor, this strategy appears simply as prone to set off a semantic race to the underside as to attain a balanced, Wikipedia-style neutrality. Calibrating information curation round consumer demand dangers amplifying the preferences of the lowest-common-denominator whereas marginalizing area of interest customers, since majority pursuits will inevitably carry better weight.
Nonetheless, let’s check out how the paper tackles the problem.
Distilling Ideas with Discretion
The researchers used the 2024 VidProM dataset because the supply for subject evaluation that will later inform the venture’s web-scraping.
This dataset was chosen, the authors state, as a result of it’s the solely publicly-available 1m+ dataset ‘written by actual customers’ – and it ought to be said that this dataset was itself curated by the 2 authors of the brand new paper.
The paper explains*:
‘First, we embed all 1.67 million prompts from VidProM into 384-dimensional vectors utilizing SentenceTransformers Subsequent, we cluster these vectors with Ok-means. Notice that right here we preset the variety of clusters to a comparatively massive worth, i.e., 2, 000, and merge comparable clusters within the subsequent step.
‘Lastly, for every cluster, we ask GPT-4o to conclude a subject [one or two words].’
The authors level out that sure ideas are distinct however notably adjoining, corresponding to church and cathedral. Too granular a standards for instances of this type would result in idea embeddings (as an illustration) for every kind of canine breed, as an alternative of the time period canine; whereas too broad a standards may corral an extreme variety of sub-concepts right into a single over-crowded idea; due to this fact the paper notes the balancing act needed to guage such instances.
Singular and plural varieties had been merged, and verbs restored to their base (infinitive) varieties. Excessively broad phrases – corresponding to animation, scene, movie and motion – had been eliminated.
Thus 1,291 matters had been obtained (with the total record obtainable within the supply paper’s supplementary part).
Choose Internet-Scraping
Subsequent, the researchers used the official YouTube API to hunt movies primarily based on the factors distilled from the 2024 dataset, in search of to acquire 500 movies for every subject. Moreover the requisite artistic commons license, every video needed to have a decision of 720p or greater, and needed to be shorter than 4 minutes.
On this manner 586,490 movies had been scraped from YouTube.
The authors in contrast the YouTube ID of the downloaded movies to a lot of standard datasets: OpenVid-1M; HD-VILA-100M; Intern-Vid; Koala-36M; LVD-2M; MiraData; Panda-70M; VidGen-1M; and WebVid-10M.
They discovered that just one,675 IDs (the aforementioned 0.29%) of the VideoUFO clips featured in these older collections, and it needs to be conceded that whereas the dataset comparability record shouldn’t be exhaustive, it does embrace all the most important and most influential gamers within the generative video scene.
Splits and Evaluation
The obtained movies had been subsequently segmented into a number of clips, in response to the methodology outlined within the Panda-70M paper cited above. Shot boundaries had been estimated, assemblies stitched, and the concatenated movies divided into single clips, with transient and detailed captions supplied.
Every information entry within the VideoUFO dataset includes a clip, an ID, begin and finish occasions, and a short and an in depth caption.
The transient captions had been dealt with by the Panda-70M methodology, and the detailed video captions by Qwen2-VL-7B, alongside the rules established by Open-Sora-Plan. In instances the place clips didn’t efficiently embody the supposed goal idea, the detailed captions for every such clip had been fed into GPT-4o mini, with a view to confirm whether or not it was actually a match for the subject. Although the authors would have most popular analysis by way of GPT-4o, this is able to have been too costly for thousands and thousands of video clips.
Video high quality evaluation was dealt with with six strategies from the VBench venture .
Comparisons
The authors repeated the subject extraction course of on the aforementioned prior datasets. For this, it was essential to semantically-match the derived classes of VideoUFO to the inevitably completely different classes within the different collections; it needs to be conceded that such processes provide solely approximated equal classes, and due to this fact this can be too subjective a course of to vouchsafe empirical comparisons.
Nonetheless, within the picture beneath we see the outcomes the researchers obtained by this methodology:
Comparability of the basic attributes derived throughout VideoUFO and the prior datasets.
The researchers acknowledge that their evaluation relied on the present captions and descriptions supplied in every dataset. They admit that re-captioning older datasets utilizing the identical methodology as VideoUFO may have provided a extra direct comparability. Nonetheless, given the sheer quantity of information factors, their conclusion that this strategy could be prohibitively costly appears justified.
Technology
The authors developed a benchmark to guage text-to-video fashions’ efficiency on user-focused ideas, titled BenchUFO. This entailed choosing 791 nouns from the 1,291 distilled consumer matters in VideoUFO. For every chosen subject, ten textual content prompts from VidProM had been then randomly chosen.
Every immediate was handed by way of to a text-to-video mannequin, with the aforementioned Qwen2-VL-7B captioner used to guage the generated outcomes. With all generated movies thus captioned, SentenceTransformers was used to calculate cosine similarity for each the enter immediate and output (inferred) description in every case.
Schema for the BenchUFO course of.
The evaluated generative fashions had been: Mira; Show-1; LTX-Video; Open-Sora-Plan; Open Sora; TF-T2V; Mochi-1; HiGen; Pika; RepVideo; T2V-Zero; CogVideoX; Latte-1; Hunyuan Video; LaVie; and Pyramidal.
Moreover VideoUFO, MVDiT-VidGen and MVDit-OpenVid had been the choice coaching datasets.
The outcomes think about the Tenth-Fiftieth worst-performing and best-performing matters throughout the architectures and datasets.
Outcomes for the efficiency of public T2V fashions vs. the authors’ skilled fashions, on BenchUFO.
Right here the authors remark:
‘Present text-to-video fashions don’t persistently carry out properly throughout all user-focused matters. Particularly, there’s a rating distinction starting from 0.233 to 0.314 between the top-10 and low-10 matters. These fashions could not successfully perceive matters corresponding to “big squid”, “animal cell”, “Van Gogh”, and “historic Egyptian” resulting from inadequate coaching on such movies.
‘Present text-to-video fashions present a sure diploma of consistency of their best-performing matters. We uncover that almost all text-to-video fashions excel at producing movies on animal-related matters, corresponding to ‘seagull’, ‘panda’, ‘dolphin’, ‘camel’, and ‘owl’. We infer that that is partly resulting from a bias in the direction of animals in present video datasets.’
Conclusion
VideoUFO is an impressive providing if solely from the standpoint of recent information. If there was no error in evaluating and eliminating YouTube IDs, and if the dataset comprises a lot materials that’s new to the analysis scene, it’s a uncommon and doubtlessly helpful proposition.
The draw back is that one wants to provide credence to the core methodology; for those who do not imagine that consumer demand ought to inform web-scraping formulation, you would be shopping for right into a dataset that comes with its personal units of troubling biases.
Additional, the utility of the distilled matters relies on each the reliability of the distilling methodology used (which is mostly hampered by funds constraints), and in addition the formulation strategies for the 2024 dataset that gives the supply materials.
That mentioned, VideoUFO actually deserves additional investigation – and it’s available at Hugging Face.
* My substitution of the authors’ citations for hyperlinks.
First printed Wednesday, March 5, 2025