7 papers · 1 filter
What Comes Next? Omni-StoryBench for Evaluating Story-Grounded Omnimodal Generation
Sieun Hyeon, Yejoon Lee, Mintaek Lim +3
Omnimodal evaluation should go beyond independent text, image, and speech production: individually plausible outputs may not express a coherent shared event. We introduce Omni-Stor…
Beyond Retraining-Free MoE Compression: A Cost-Normalized Study of Post-Compression Adjustment
Sieun Hyeon, Jaeyoung Do
Retraining-free MoE compression reduces deployment memory by pruning or merging experts, but often treats the compressed checkpoint as the final artifact. We argue that this view i…
CIPHER: Counterfeit Image Pattern High-level Examination via Representation
Kyeonghun Kim, Youngung Han, Seoyoung Ju +9
The rapid progress of generative adversarial networks (GANs) and diffusion models has enabled the creation of synthetic faces that are increasingly difficult to distinguish from re…
Dynin-Omni: Omnimodal Unified Large Diffusion Language Model
Jaeik Kim, Woojin Kim, Jihwan Hong +8
We present Dynin-Omni, the first masked-diffusion-based omnimodal foundation model that unifies text, image, and speech understanding and generation, together with video understand…
Is Retraining-Free Enough? The Necessity of Router Calibration for Efficient MoE Compression
Sieun Hyeon, Jaeyoung Do
Mixture-of-Experts (MoE) models scale capacity efficiently, but their massive parameter footprint creates a deployment-time memory bottleneck. We organize retraining-free MoE compr…
MATA: Multi-Agent Framework for Reliable and Flexible Table Question Answering
Sieun Hyeon, Jusang Oh, Sunghwan Steve Cho +1
Recent advances in Large Language Models (LLMs) have significantly improved table understanding tasks such as Table Question Answering (TableQA), yet challenges remain in ensuring…
VALUEFLOW: Toward Pluralistic and Steerable Value-based Alignment in Large Language Models
Woojin Kim, Sieun Hyeon, Jusang Oh +1
Aligning Large Language Models (LLMs) with the diverse spectrum of human values remains a central challenge: preference-based methods often fail to capture deeper motivational prin…