Related papers: The Multi-Branched Method of Moments for Queueing …

Long-context language models exhibit impressive performance but remain challenging to deploy due to high GPU memory demands during inference. We propose Memory-efficient Offloaded Mini-sequence Inference (MOM), a method that partitions…

Queueing networks are systems of theoretical interest that find widespread use in the performance evaluation of interconnected resources. In comparison to counterpart models in genetics or mathematical biology, the stochastic (jump)…

Although deep learning has demonstrated remarkable capability in learning from unstructured data, modern tree-based ensemble models remain superior in extracting relevant information and learning from structured datasets. While several…

The traditional RAG paradigm, which typically engages in the comprehension of relevant text chunks in response to received queries, inherently restricts both the depth of knowledge internalization and reasoning capabilities. To address this…

In network quality of service provisioning, premium services generally require to keep a very small loss probability, which is infeasible to measure directly. The proposed virtual queue scheme estimates the small packet loss probability of…

By exploiting the superiority of non-orthogonal multiple access (NOMA), NOMA-aided mobile edge computing (MEC) can provide scalable and low-latency computing services for the Internet of Things. However, given the prevalent stochasticity of…

Deep neural networks conventionally employ end-to-end backpropagation for their training process, which lacks biological credibility and triggers a locking dilemma during network parameter updates, leading to significant GPU memory use.…

Active Queue Management (AQM), a network-layer congestion control technique endorsed by the Internet Engineering Task Force (IETF), encourages routers to discard packets before the occurrence of buffer overflow. Traditional AQM techniques…

The Class Activation Map (CAM) lookup of a neural network tells us to which regions the neural network focuses when it makes a decision. In the past, the CAM search method was dependent upon a specific internal module of the network. It has…

Quasi-Monte Carlo (qMC) methods are a powerful alternative to classical Monte-Carlo (MC) integration. Under certain conditions, they can approximate the desired integral at a faster rate than the usual Central Limit Theorem, resulting in…

External memory systems are pivotal for enabling Large Language Model (LLM) agents to maintain persistent knowledge and perform long-horizon decision-making. Existing paradigms typically follow a two-stage process: computationally expensive…

The emergence of new applications brings multi-class traffic with diverse quality of service (QoS) requirements to wide area networks (WANs), motivating research in traffic engineering (TE). In recent years, novel centralized and…

Precise pointer analysis is a foundational component of many client analyses and optimizations. Scaling flow- and context-sensitive pointer analysis has been a long-standing challenge, suffering from combinatorial growth in both memory…

Is it always necessary to compute tokens from shallow to deep layers in Transformers? The continued success of vanilla Transformers and their variants suggests an undoubted "yes". In this work, however, we attempt to break the depth-ordered…

As long-context language modeling becomes increasingly important, the cost of maintaining and attending to large Key/Value (KV) caches grows rapidly, becoming a major bottleneck in both training and inference. While prior works such as…

With data sizes constantly expanding, and with classical machine learning algorithms that analyze such data requiring larger and larger amounts of computation time and storage space, the need to distribute computation and memory…

A novel non-orthogonal multiple access (NOMA) based cache-aided mobile edge computing (MEC) framework is proposed. For the purpose of efficiently allocating communication and computation resources to users' computation tasks requests, we…

Transformers have achieved remarkable success in sequence modeling and beyond but suffer from quadratic computational and memory complexities with respect to the length of the input sequence. Leveraging techniques include sparse and linear…

Inference networks of traditional Variational Autoencoders (VAEs) are typically amortized, resulting in relatively inaccurate posterior approximation compared to instance-wise variational optimization. Recent semi-amortized approaches were…

Retrieval-augmented generation (RAG) has emerged as a popular approach to steering the output of a large language model (LLM) by incorporating retrieved contexts as inputs. However, existing work observed the generator bias, such that…