Min‐Soo Kim

Minsoo Kim, Kyuhong Shim, Jungwook Choi et al.

Modern multimodal large language models (MLLMs) can reason over hour-long video, yet their key-value (KV) cache grows linearly with time-quickly exceeding the fixed memory of phones, AR glasses, and edge robots. Prior compression schemes either assume the whole video and user query are available offline or must first build the full cache, so memory still scales with stream length. InfiniPot-V is the first training-free, query-agnostic framework that enforces a hard, length-independent memory cap for streaming video understanding. During video encoding it monitors the cache and, once a user-set threshold is reached, runs a lightweight compression pass that (i) removes temporally redundant tokens via Temporal-axis Redundancy (TaR) metric and (ii) keeps semantically significant tokens via Value-Norm (VaN) ranking. Across four open-source MLLMs and four long-video and streaming-video benchmarks, InfiniPot-V cuts peak GPU memory by up to 94%, sustains real-time generation, and matches or surpasses full-cache accuracy-even in multi-turn dialogues. By dissolving the KV cache bottleneck without retraining or query knowledge, InfiniPot-V closes the gap for on-device streaming video assistants.

Minsoo Kim, Jip Kim

Optimal transmission switching (OTS) improves optimal power flow (OPF) by selectively opening transmission lines, but its mixed-integer formulation increases computational complexity, especially on large grids. To deal with this, we propose a dispatch-aware deep neural network (DA-DNN) that accelerates DC-OTS without relying on pre-solved labels. DA-DNN predicts line states and passes them through a differentiable DC-OPF layer, using the resulting generation cost as the loss function so that all physical network constraints are enforced throughout training and inference. In addition, we adopt a customized weight-bias initialization that keeps every forward pass feasible from the first iteration, which allows stable learning on large grids. Once trained, the proposed DA-DNN produces a provably feasible topology and dispatch pair in the same time as solving the DCOPF, whereas conventional mixed-integer solvers become intractable. As a result, the proposed method successfully captures the economic advantages of OTS while maintaining scalability.