Concept Weaver: Enabling Multi-Concept Fusion in Text-to-Image Models

While there has been significant progress in customizing text-to-image generation models, generating images that combine multiple personalized concepts remains challenging. In this work, we introduce Concept Weaver, a method for composing customized text-to-image diffusion models at inference time. Specifically, the method breaks the process into two steps: creating a template image aligned with the semantics of input prompts, and then personalizing the template using a concept fusion strategy. The fusion strategy incorporates the appearance of the target concepts into the template image while retaining its structural details. The results indicate that our method can generate multiple custom concepts with higher identity fidelity compared to alternative approaches. Furthermore, the method is shown to seamlessly handle more than two concepts and closely follow the semantic meaning of the input prompt without blending appearances across different subjects.

Next Steps for Human-Centered Generative AI: A Technical Perspective

Through iterative, cross-disciplinary discussions, we define and propose next-steps for Human-centered Generative AI (HGAI) from a technical perspective. We contribute a roadmap that lays out future directions of Generative AI spanning three levels: Aligning with human values; Accommodating humans' expression of intents; and Augmenting humans' abilities in a collaborative workflow. This roadmap intends to draw interdisciplinary research teams to a comprehensive list of emergent ideas in HGAI, identifying their interested topics while maintaining a coherent big picture of the future work landscape.

ReelFramer: Co-creating News Reels on Social Media with Generative AI

Short videos on social media are a prime way many young people find and consume content. News outlets would like to reach audiences through news reels, but currently struggle to translate traditional journalistic formats into the short, entertaining videos that match the style of the platform. There are many ways to frame a reel-style narrative around a news story, and selecting one is a challenge. Different news stories call for different framings, and require a different trade-off between entertainment and information. We present a system called ReelFramer that uses text and image generation to help journalists explore multiple narrative framings for a story, then generate scripts, character boards and storyboards they can edit and iterate on. A user study of five graduate students in journalism-related fields found the system greatly eased the burden of transforming a written story into a reel, and that exploring framings to find the right one was a rewarding process.

Audio-driven Neural Gesture Reenactment with Video Motion Graphs

Human speech is often accompanied by body gestures including arm and hand gestures. We present a method that reenacts a high-quality video with gestures matching a target speech audio. The key idea of our method is to split and re-assemble clips from a reference video through a novel video motion graph encoding valid transitions between clips. To seamlessly connect different clips in the reenactment, we propose a pose-aware video blending network which synthesizes video frames around the stitched frames between two clips. Moreover, we developed an audio-based gesture searching algorithm to find the optimal order of the reenacted frames. Our system generates reenactments that are consistent with both the audio rhythms and the speech content. We evaluate our synthesized video quality quantitatively, qualitatively, and with user studies, demonstrating that our method produces videos of much higher quality and consistency with the target audio compared to previous work and baselines.

Audio-Visual Fusion Layers for Event Type Aware Video Recognition

Human brain is continuously inundated with the multisensory information and their complex interactions coming from the outside world at any given moment. Such information is automatically analyzed by binding or segregating in our brain. While this task might seem effortless for human brains, it is extremely challenging to build a machine that can perform similar tasks since complex interactions cannot be dealt with single type of integration but requires more sophisticated approaches. In this paper, we propose a new model to address the multisensory integration problem with individual event-specific layers in a multi-task learning scheme. Unlike previous works where single type of fusion is used, we design event-specific layers to deal with different audio-visual relationship tasks, enabling different ways of audio-visual formation. Experimental results show that our event-specific layers can discover unique properties of the audio-visual relationships in the videos. Moreover, although our network is formulated with single labels, it can output additional true multi-labels to represent the given videos. We demonstrate that our proposed framework also exposes the modality bias of the video data category-wise and dataset-wise manner in popular benchmark datasets.

Can one hear the shape of a neural network?: Snooping the GPU via Magnetic Side Channel

Neural network applications have become popular in both enterprise and personal settings. Network solutions are tuned meticulously for each task, and designs that can robustly resolve queries end up in high demand. As the commercial value of accurate and performant machine learning models increases, so too does the demand to protect neural architectures as confidential investments. We explore the vulnerability of neural networks deployed as black boxes across accelerated hardware through electromagnetic side channels. We examine the magnetic flux emanating from a graphics processing unit's power cable, as acquired by a cheap $3 induction sensor, and find that this signal betrays the detailed topology and hyperparameters of a black-box neural network model. The attack acquires the magnetic signal for one query with unknown input values, but known input dimensions. The network reconstruction is possible due to the modular layer sequence in which deep neural networks are evaluated. We find that each layer component's evaluation produces an identifiable magnetic signal signature, from which layer topology, width, function type, and sequence order can be inferred using a suitably trained classifier and a joint consistency optimization based on integer programming. We study the extent to which network specifications can be recovered, and consider metrics for comparing network similarity. We demonstrate the potential accuracy of this side channel attack in recovering the details for a broad range of network architectures, including random designs. We consider applications that may exploit this novel side channel exposure, such as adversarial transfer attacks. In response, we discuss countermeasures to protect against our method and other similar snooping techniques.

Unified Multisensory Perception: Weakly-Supervised Audio-Visual Video Parsing

In this paper, we introduce a new problem, named audio-visual video parsing, which aims to parse a video into temporal event segments and label them as either audible, visible, or both. Such a problem is essential for a complete understanding of the scene depicted inside a video. To facilitate exploration, we collect a Look, Listen, and Parse (LLP) dataset to investigate audio-visual video parsing in a weakly-supervised manner. This task can be naturally formulated as a Multimodal Multiple Instance Learning (MMIL) problem. Concretely, we propose a novel hybrid attention network to explore unimodal and cross-modal temporal contexts simultaneously. We develop an attentive MMIL pooling method to adaptively explore useful audio and visual content from different temporal extent and modalities. Furthermore, we discover and mitigate modality bias and noisy label issues with an individual-guided learning mechanism and label smoothing technique, respectively. Experimental results show that the challenging audio-visual video parsing can be achieved even with only video-level weak labels. Our proposed framework can effectively leverage unimodal and cross-modal temporal contexts and alleviate modality bias and noisy labels problems.

Deep Audio Prior

Deep convolutional neural networks are known to specialize in distilling compact and robust prior from a large amount of data. We are interested in applying deep networks in the absence of training dataset. In this paper, we introduce deep audio prior (DAP) which leverages the structure of a network and the temporal information in a single audio file. Specifically, we demonstrate that a randomly-initialized neural network can be used with carefully designed audio prior to tackle challenging audio problems such as universal blind source separation, interactive audio editing, audio texture synthesis, and audio co-separation. To understand the robustness of the deep audio prior, we construct a benchmark dataset \emph{Universal-150} for universal sound source separation with a diverse set of sources. We show superior audio results than previous work on both qualitative and quantitative evaluations. We also perform thorough ablation study to validate our design choices.

Scene-Aware Audio for 360\textdegree{} Videos

Although 360\textdegree{} cameras ease the capture of panoramic footage, it remains challenging to add realistic 360\textdegree{} audio that blends into the captured scene and is synchronized with the camera motion. We present a method for adding scene-aware spatial audio to 360\textdegree{} videos in typical indoor scenes, using only a conventional mono-channel microphone and a speaker. We observe that the late reverberation of a room's impulse response is usually diffuse spatially and directionally. Exploiting this fact, we propose a method that synthesizes the directional impulse response between any source and listening locations by combining a synthesized early reverberation part and a measured late reverberation tail. The early reverberation is simulated using a geometric acoustic simulation and then enhanced using a frequency modulation method to capture room resonances. The late reverberation is extracted from a recorded impulse response, with a carefully chosen time duration that separates out the late reverberation from the early reverberation. In our validations, we show that our synthesized spatial audio matches closely with recordings using ambisonic microphones. Lastly, we demonstrate the strength of our method in several applications.

Interacting with Acoustic Simulation and Fabrication

Incorporating accurate physics-based simulation into interactive design tools is challenging. However, adding the physics accurately becomes crucial to several emerging technologies. For example, in virtual/augmented reality (VR/AR) videos, the faithful reproduction of surrounding audios is required to bring the immersion to the next level. Similarly, as personal fabrication is made possible with accessible 3D printers, more intuitive tools that respect the physical constraints can help artists to prototype designs. One main hurdle is the sheer amount of computation complexity to accurately reproduce the real-world phenomena through physics-based simulation. In my thesis research, I develop interactive tools that implement efficient physics-based simulation algorithms for automatic optimization and intuitive user interaction.