arXiv subjects

Yanxuan Yu

Publications and source records attributed to Yanxuan Yu.

26 records · Page 2Linked to original sources

arXiv2025

Cognitive Load Traces as Symbolic and Visual Accounts of Deep Model Cognition

We propose \textbf{Cognitive Load Traces} (CLTs) as a mid-level interpretability framework for deep models, inspired by Cognitive Load Theory in human cognition. CLTs are defined as symbolic, temporally varying functions that quantify model-internal resource allocation. Formally, we represent CLTs as a three-component stochastic process $(\mathrm{IL}_t, \mathrm{EL}_t, \mathrm{GL}_t)$, corresponding to \emph{Intrinsic}, \emph{Extraneous}, and \emph{Germane} load. Each component is instantiated through measurable proxies such as attention entropy, KV-cache miss ratio, representation dispersion, and decoding stability. We propose both symbolic formulations and visualization methods (load curves, simplex diagrams) that enable interpretable analysis of reasoning dynamics. Experiments on reasoning and planning benchmarks show that CLTs predict error-onset, reveal cognitive strategies, and enable load-guided interventions that improve reasoning efficiency by 15-30\% while maintaining accuracy.

cs.AI↗

arXiv2025

HSGM: Hierarchical Segment-Graph Memory for Scalable Long-Text Semantics

Semantic parsing of long documents remains challenging due to quadratic growth in pairwise composition and memory requirements. We introduce \textbf{Hierarchical Segment-Graph Memory (HSGM)}, a novel framework that decomposes an input of length $N$ into $M$ meaningful segments, constructs \emph{Local Semantic Graphs} on each segment, and extracts compact \emph{summary nodes} to form a \emph{Global Graph Memory}. HSGM supports \emph{incremental updates} -- only newly arrived segments incur local graph construction and summary-node integration -- while \emph{Hierarchical Query Processing} locates relevant segments via top-$K$ retrieval over summary nodes and then performs fine-grained reasoning within their local graphs. Theoretically, HSGM reduces worst-case complexity from $O(N^2)$ to $O\!\left(N\,k + (N/k)^2\right)$, with segment size $k \ll N$, and we derive Frobenius-norm bounds on the approximation error introduced by node summarization and sparsification thresholds. Empirically, on three benchmarks -- long-document AMR parsing, segment-level semantic role labeling (OntoNotes), and legal event extraction -- HSGM achieves \emph{2--4$\times$ inference speedup}, \emph{$>60\%$ reduction} in peak memory, and \emph{$\ge 95\%$} of baseline accuracy. Our approach unlocks scalable, accurate semantic modeling for ultra-long texts, enabling real-time and resource-constrained NLP applications.

cs.AI↗

arXiv2025

TinyServe: Query-Aware Cache Selection for Efficient LLM Serving

Serving large language models (LLMs) efficiently remains challenging due to the high memory and latency overhead of key-value (KV) cache access during autoregressive decoding. We present \textbf{TinyServe}, a lightweight and extensible serving system for deploying tiny LLMs (e.g., TinyLLaMA, GPT2-345M) with support for structured KV sparsity, plugin-based token selection, and hardware-efficient attention kernels. Unlike prior simulation frameworks, TinyServe executes real-time decoding with configurable sparsity strategies and fine-grained instrumentation. To reduce decoding cost, we introduce a \textit{query-aware page selection} mechanism that leverages bounding-box metadata to estimate attention relevance between the query and KV cache blocks. This enables selective KV loading with minimal overhead and no model modifications. Our fused CUDA kernel integrates page scoring, sparse memory access, and masked attention in a single pass. Experiments show that TinyServe achieves up to \textbf{3.4x} speedup and over \textbf{2x} memory savings with negligible accuracy drop. Additional analysis of cache reuse, page hit rate, and multi-GPU scaling confirms its practicality as an efficient system-level design for LLM training and inference research on resource-constrained hardware.

cs.DC↗

arXiv2025

SemToken: Semantic-Aware Tokenization for Efficient Long-Context Language Modeling

Tokenization plays a critical role in language modeling, yet existing approaches such as Byte-Pair Encoding (BPE) or WordPiece operate purely on frequency statistics, ignoring the underlying semantic structure of text. This leads to over-tokenization of semantically redundant spans and underutilization of contextual coherence, particularly in long-context scenarios. In this work, we propose \textbf{SemToken}, a semantic-aware tokenization framework that jointly reduces token redundancy and improves computation efficiency. SemToken first extracts contextual semantic embeddings via lightweight encoders and performs local semantic clustering to merge semantically equivalent tokens. Then, it allocates heterogeneous token granularity based on semantic density, allowing finer-grained tokenization in content-rich regions and coarser compression in repetitive or low-entropy spans. SemToken can be seamlessly integrated with modern language models and attention acceleration methods. Experiments on long-context language modeling benchmarks such as WikiText-103 and LongBench show that SemToken achieves up to $2.4\times$ reduction in token count and $1.9\times$ speedup, with negligible or no degradation in perplexity and downstream accuracy. Our findings suggest that semantic structure offers a promising new axis for optimizing tokenization and computation in large language models.

cs.CL↗

arXiv2025

QuickMerge++: Fast Token Merging with Autoregressive Prior

As generative models scale to larger inputs across language, vision, and video domains, the cost of token-level computation has become a key bottleneck. While prior work suggests that only a subset of tokens significantly influence downstream predictions, most token selection methods are static, modality-specific, or incompatible with autoregressive generation. In this paper, we propose QuickMerge, a lightweight token merging framework designed for efficient next-token prediction. QuickMerge dynamically selects a reduced number of tokens based on attention norm magnitude, guided by an entropy-based budget estimator. To preserve autoregressive compatibility, we introduce a lightweight transformer prior trained over the merged token sequence. By combining semantic salience estimation, flexible token budgets, and AR alignment, QuickMerge enables accurate generation with fewer tokens. We evaluate QuickMerge across multi-modality domains, demonstrating consistent improvements in compute-accuracy tradeoffs. Specifically, QuickMerge reduces token counts sustantially while matching as well as exceeding the performance of learned tokenizers and fixed-patch baselines.

cs.AI↗

arXiv2025

MT2ST: Adaptive Multi-Task to Single-Task Learning

Efficient machine learning (ML) has become increasingly important as models grow larger and data volumes expand. In this work, we address the trade-off between generalization in multi-task learning (MTL) and precision in single-task learning (STL) by introducing the Multi-Task to Single-Task (MT2ST) framework. MT2ST is designed to enhance training efficiency and accuracy in multi-modal tasks, showcasing its value as a practical application of efficient ML.

cs.LG↗

arXiv2025

Designing Large Foundation Models for Efficient Training and Inference: A Survey

This paper focuses on modern efficient training and inference technologies on foundation models and illustrates them from two perspectives: model and system design. Model and System Design optimize LLM training and inference from different aspects to save computational resources, making LLMs more efficient, affordable, and more accessible. The paper list repository is available at https://github.com/NoakLiu/Efficient-Foundation-Models-Survey.

cs.DC↗

arXiv2025

MobileNetV2: A lightweight classification model for home-based sleep apnea screening

This study proposes a novel lightweight neural network model leveraging features extracted from electrocardiogram (ECG) and respiratory signals for early OSA screening. ECG signals are used to generate feature spectrograms to predict sleep stages, while respiratory signals are employed to detect sleep-related breathing abnormalities. By integrating these predictions, the method calculates the apnea-hypopnea index (AHI) with enhanced accuracy, facilitating precise OSA diagnosis. The method was validated on three publicly available sleep apnea databases: the Apnea-ECG database, the UCDDB dataset, and the MIT-BIH Polysomnographic database. Results showed an overall OSA detection accuracy of 0.978, highlighting the model's robustness. Respiratory event classification achieved an accuracy of 0.969 and an area under the receiver operating characteristic curve (ROC-AUC) of 0.98. For sleep stage classification, in UCDDB dataset, the ROC-AUC exceeded 0.85 across all stages, with recall for Sleep reaching 0.906 and specificity for REM and Wake states at 0.956 and 0.937, respectively. This study underscores the potential of integrating lightweight neural networks with multi-signal analysis for accurate, portable, and cost-effective OSA screening, paving the way for broader adoption in home-based and wearable health monitoring systems.

cs.LG↗