Related papers: ODMA: On-Demand Memory Allocation Strategy for LLM…
Exploring channel dimensions has been the driving force behind breakthroughs in successive generations of mobile communication systems. In 5G, space division multiple access (SDMA) leveraging massive MIMO has been crucial in enhancing…
The Key-Value (KV) cache is a crucial component in serving transformer-based autoregressive large language models (LLMs), enabling faster inference by storing previously computed KV vectors. However, its memory consumption scales linearly…
Low-Rank Adaptation (LoRA) enables efficient fine-tuning of large language models but suffers from catastrophic forgetting when learned updates interfere with the dominant singular directions that encode essential pre-trained knowledge. We…
Memory allocation, though constituting only a small portion of the executed code, can have a "butterfly effect" on overall program performance, leading to significant and far-reaching impacts. Despite accounting for just approximately 5% of…
Mixture-of-Experts (MoE), while offering significant advantages as a Large Language Model (LLM) architecture, faces substantial challenges when deployed on low-cost edge devices with tight memory constraints. Expert offloading mitigates…
Vision-language models (VLMs) struggle in open-world applications, where out-of-distribution (OOD) concepts often trigger cross-modal alignment collapse and severely degrade zero-shot performance. We identify the root cause as modal…
Orthogonal Frequency Division Multiple Access (OFDMA) as well as other orthogonal multiple access techniques fail to achieve the system capacity limit in the uplink due to the exclusivity in resource allocation. This issue is more prominent…
Contention-based wireless channel access methods like CSMA and ALOHA paved the way for the rise of the Internet of Things in industrial applications (IIoT). However, to cope with increasing demands for reliability and throughput, several…
In this paper, we propose a fast decoder algorithm for uniquely decodable (errorless) code sets for overloaded synchronous optical code-division multiple-access (O-CDMA) systems. The proposed decoder is designed in a such a way that the…
Management of disk scheduling is a very important aspect of operating system. Performance of the disk scheduling completely depends on how efficient is the scheduling algorithm to allocate services to the request in a better manner. Many…
On-device deployment of Large Language Models (LLMs) frequently leverages Low-Rank Adapters (LoRAs) to support diverse downstream tasks under tight resource constraints. To address the limited storage capacity of mobile devices, recent…
Attention-based Transformers have revolutionized natural language processing (NLP) and shown strong performance in computer vision (CV) tasks. However, as the input sequence varies, the computational bottlenecks in Transformer models…
Large Language Models (LLMs) can perform zero-shot learning on unseen tasks and few-shot learning on complex reasoning tasks. However, resource-limited mobile edge networks struggle to support long-context LLM serving for LLM agents during…
Future networks are expected to connect an enormous number of nodes wirelessly using wide-band transmission. This brings great challenges. To avoid collecting a large amount of data from the massive number of nodes, computation over…
Delay-Doppler (DD) signal processing has emerged as a powerful tool for analyzing multipath and time-varying channel effects. Due to the inherent sparsity of the wireless channel in the DD domain, compressed sensing (CS) based techniques,…
In this letter, we investigate the resource allocation for downlink multi-cell coordinated OFDMA wireless networks, in which power allocation and subcarrier scheduling are jointly optimized. Aiming at maximizing the weighted sum of the…
This work studies how to adaptively recompute key-value (KV) caches for diffusion large language models (DLMs) to maximize prediction accuracy while minimizing decoding latency. Prior methods' decoders recompute QKV for all tokens at every…
In order to meet the ever-increasing demand for high throughput in WiFi networks, the IEEE 802.11ax (11ax) standard introduces orthogonal frequency division multiple access (OFDMA). In this letter, we address the station-resource unit…
This paper focuses on online kernel learning over a decentralized network. Each agent in the network receives continuous streaming data locally and works collaboratively to learn a nonlinear prediction function that is globally optimal in…
Modern online large language model (LLM) services, such as Retrieval-Augmented Generation (RAG) and agent systems, increasingly expose two prominent characteristics: prompt segmentation (e.g., system instructions, retrieved passages, tool…