←
Tiffany Tseng 🍡11,216 posts
Tiffany Tseng 🍡 avatar

Tiffany Tseng 🍡

@scientiffic

designing tools for making. assistant professor of computer science @barnardcollege. mostly at @scientiffic@mas.to

◷ Joined October 2009▧ 3,286 Photos and videos
Tiffany Tseng 🍡 avatar
Tiffany Tseng 🍡@scientiffic·16 May 2024

I'm so happy to share that I'm joining @BarnardCollege @Columbia as an Assistant Professor of Computer Science starting in July 2024! Excited and thankful for this opportunity to start my own lab developing design tools for creativity and learning. 😊 See you back in NYC! 🏙️

Post attachment
Tiffany Tseng 🍡 avatar
Tiffany Tseng 🍡@scientiffic·9 Mar 2024

excited to share a demo video of how our Keyframer prototype works! users input their own SVG images and enter prompts to animate them. because animations are created with code, you can edit generated designs directly for greater creative control. machinelearning.apple.com/re…

1:52
Regina Cheng (she/her) avatar
⟳ Tiffany Tseng 🍡 retweeted
Regina Cheng (she/her)@ReginaRCheng·12 Feb 2024

Our preprint paper is featured by @huggingface! Also: arxiv.org/pdf/2402.06071.pdf We present Keyframer, a LLM-based system for users to create & iterate on animation designs 🎨 This is part of my ongoing residency at @Apple - so grateful to collaborate w @scientiffic & @jwnichls 💯

AK@_akhaliq12 Feb 2024

Apple presents Keyframer Empowering Animation Design using Large Language Models paper page: huggingface.co/papers/2402.0… Large language models (LLMs) have the potential to impact a wide range of creative domains, but the application of LLMs to animation is underexplored and presents novel challenges such as how users might effectively describe motion in natural language. In this paper, we present Keyframer, a design tool for animating static images (SVGs) with natural language. Informed by interviews with professional animation designers and engineers, Keyframer supports exploration and refinement of animations through the combination of prompting and direct editing of generated output. The system also enables users to request design variants, supporting comparison and ideation. Through a user study with 13 participants, we contribute a characterization of user prompting strategies, including a taxonomy of semantic prompt types for describing motion and a 'decomposed' prompting style where users continually adapt their goals in response to generated output.We share how direct editing along with prompting enables iteration beyond one-shot prompting interfaces common in generative tools today. Through this work, we propose how LLMs might empower a range of audiences to engage with animation creation.

Post attachment
Tiffany Tseng 🍡 avatar
Tiffany Tseng 🍡@scientiffic·12 Feb 2024

🌟here's a pre-print of our new paper applying LLMs to animation design! our system Keyframer lets people iteratively animate static images using natural language + direct editors. arxiv.org/pdf/2402.06071.pdf awesome collaborating with @ReginaRCheng and @jwnichls on this project 😊

AK@_akhaliq12 Feb 2024

Apple presents Keyframer Empowering Animation Design using Large Language Models paper page: huggingface.co/papers/2402.0… Large language models (LLMs) have the potential to impact a wide range of creative domains, but the application of LLMs to animation is underexplored and presents novel challenges such as how users might effectively describe motion in natural language. In this paper, we present Keyframer, a design tool for animating static images (SVGs) with natural language. Informed by interviews with professional animation designers and engineers, Keyframer supports exploration and refinement of animations through the combination of prompting and direct editing of generated output. The system also enables users to request design variants, supporting comparison and ideation. Through a user study with 13 participants, we contribute a characterization of user prompting strategies, including a taxonomy of semantic prompt types for describing motion and a 'decomposed' prompting style where users continually adapt their goals in response to generated output.We share how direct editing along with prompting enables iteration beyond one-shot prompting interfaces common in generative tools today. Through this work, we propose how LLMs might empower a range of audiences to engage with animation creation.

Post attachment
Jeffrey Nichols avatar
⟳ Tiffany Tseng 🍡 retweeted
Jeffrey Nichols@jwnichls·12 Feb 2024

I'm biased, but I think this is pretty great work led by @scientiffic and @ReginaRCheng! We apply a human-centered design process to investigate how LLMs can be used by designers of all experience levels to create animations.

AK@_akhaliq12 Feb 2024

Apple presents Keyframer Empowering Animation Design using Large Language Models paper page: huggingface.co/papers/2402.0… Large language models (LLMs) have the potential to impact a wide range of creative domains, but the application of LLMs to animation is underexplored and presents novel challenges such as how users might effectively describe motion in natural language. In this paper, we present Keyframer, a design tool for animating static images (SVGs) with natural language. Informed by interviews with professional animation designers and engineers, Keyframer supports exploration and refinement of animations through the combination of prompting and direct editing of generated output. The system also enables users to request design variants, supporting comparison and ideation. Through a user study with 13 participants, we contribute a characterization of user prompting strategies, including a taxonomy of semantic prompt types for describing motion and a 'decomposed' prompting style where users continually adapt their goals in response to generated output.We share how direct editing along with prompting enables iteration beyond one-shot prompting interfaces common in generative tools today. Through this work, we propose how LLMs might empower a range of audiences to engage with animation creation.

Post attachment
Tiffany Tseng 🍡 avatar
Tiffany Tseng 🍡@scientiffic·8 Jun 2023

excited to share a new project! we've developed Co-ML: a collaborative experience for building machine learning models on iPads. we show how families can learn about ML together using Co-ML in a playful at-home activity. pre-print for @IDC_ACM 2023: machinelearning.apple.com/re…

0:57