Eunhwan Park

Donghoon Han, Eunhwan Park, Gisang Lee et al.

The rapid expansion of multimedia content has made accurately retrieving relevant videos from large collections increasingly challenging. Recent advancements in text-video retrieval have focused on cross-modal interactions, large-scale foundation model training, and probabilistic modeling, yet often neglect the crucial user perspective, leading to discrepancies between user queries and the content retrieved. To address this, we introduce MERLIN (Multimodal Embedding Refinement via LLM-based Iterative Navigation), a novel, training-free pipeline that leverages Large Language Models (LLMs) for iterative feedback learning. MERLIN refines query embeddings from a user perspective, enhancing alignment between queries and video content through a dynamic question answering process. Experimental results on datasets like MSR-VTT, MSVD, and ActivityNet demonstrate that MERLIN substantially improves Recall@1, outperforming existing systems and confirming the benefits of integrating LLMs into multimodal retrieval systems for more responsive and context-aware multimedia retrieval.

Junghwan Lim, Wai Ting Cheung, Minsu Ha et al.

Training strong video generation models usually requires massive datasets, large parameter counts, and substantial compute. In this work, we ask whether strong text-to-video quality is possible at a much smaller budget: fewer than 10M clips and less than 100,000 H200 GPU hours. Our core claim is that part of the answer lies in how model capacity is organized, not only in how much of it is used. In video generation, prompt alignment, temporal consistency, and fine-detail recovery can interfere with one another when they are handled through the same pathway. Motif-Video 2B addresses this by separating these roles architecturally, rather than relying on scale alone. The model combines two key ideas. First, Shared Cross-Attention strengthens text control when video token sequences become long. Second, a three-part backbone separates early fusion, joint representation learning, and detail refinement. To make this design effective under a limited compute budget, we pair it with an efficient training recipe based on dynamic token routing and early-phase feature alignment to a frozen pretrained video encoder. Our analysis shows that later blocks develop clearer cross-frame attention structure than standard single-stream baselines. On VBench, Motif-Video~2B reaches 83.76\%, surpassing Wan2.1 14B while using 7$\times$ fewer parameters and substantially less training data. These results suggest that careful architectural specialization, combined with an efficiency-oriented training recipe, can narrow or exceed the quality gap typically associated with much larger video models.

Donghoon Han, Eunhwan Park, Seunghyeon Seo

Dense image retrieval is accurate but offers limited interpretability and attribution, and it can be compute-intensive at scale. We present \textbf{BM25-V}, which applies Okapi BM25 scoring to sparse visual-word activations from a Sparse Auto-Encoder (SAE) on Vision Transformer patch features. Across a large gallery, visual-word document frequencies are highly imbalanced and follow a Zipfian-like distribution, making BM25's inverse document frequency (IDF) weighting well suited for suppressing ubiquitous, low-information words and emphasizing rare, discriminative ones. BM25-V retrieves high-recall candidates via sparse inverted-index operations and serves as an efficient first-stage retriever for dense reranking. Across seven benchmarks, BM25-V achieves Recall@200 $\geq$ 0.993, enabling a two-stage pipeline that reranks only $K{=}200$ candidates per query and recovers near-dense accuracy within $0.2$\% on average. An SAE trained once on ImageNet-1K transfers zero-shot to seven fine-grained benchmarks without fine-tuning, and BM25-V retrieval decisions are attributable to specific visual words with quantified IDF contributions.

Kun-Hui Lee, Eunhwan Park, Donghoon Han et al.

Large Language Models (LLMs) excel across a variety of language tasks yet are constrained by limited input lengths and high computational costs. Existing approaches\textemdash such as relative positional encodings (e.g., RoPE, ALiBi) and sliding window mechanisms\textemdash partially alleviate these issues but often require additional training or suffer from performance degradation with longer inputs. In this paper, we introduce \textbf{\textit{CacheFocus}}, a method that enhances length normalization and reduces inference latency without any further training. Our approach leverages query-independent, offline caching to efficiently reuse a Context KV Cache Store. We address the amplification of abnormal token distributions problem by re-positioning cached keys and introducing Layer-Adaptive Cache Pruning to discard low-relevance caches during pre-filling. Additionally, our Adaptive Positional Allocation Strategy dynamically reassigns cache positions to maximize the use of the available positional encoding range. Experiments on the Natural Questions and TriviaQA datasets demonstrate that CacheFocus outperforms alternative methods even when inputs exceed the $4$K limit of the \texttt{LLaMA-2} model, emphasizing its practical effectiveness for long-context LLMs. Moreover, even with large maximum input length of \texttt{Qwen2}, the performance of CacheFocus shows that it maintains consistent performance even as the number of documents increases, effectively managing long-text generation without degradation.