arXiv subjects

Stephen Xia

Publications and source records attributed to Stephen Xia.

15 recordsLinked to original sources

arXiv2026

PELM: Power Efficient On-Device LLM Inference with Speculative Decoding and Dynamic Voltage Frequency Scaling

Deploying Large Language Models (LLMs) directly on mobile platforms at the edge is gaining traction due to a myriad of benefits, such as increased privacy, personalization, and reduced latency. However, LLMs have heavy computational requirements, which are difficult for resource-constrained mobile and edge platforms to fulfill. In addition to limited compute resources, mobile and edge systems often have a compact form factor and lack physical mechanisms to dissipate heat generated from high processor usage rates (e.g., fans) to prevent throttling and reduced processing power, which LLMs can easily cause. To mitigate these effects, prior works have proposed various power governing strategies, such as dynamic voltage and frequency scaling (DVFS), for reducing power and heat generation for heavy computational tasks on mobile platforms. Recently, DVFS methods tailored for mobile LLMs have also been proposed. However, these methods mostly focus on optimizing hardware parameters and processor frequencies, and they fall short under some thermally constrained scenarios. Drawing from recent advances in machine learning, we identify and take advantage of the key insight that not all tokens require full-depth inference to maintain high-quality generation. Motivated by this, we present PELM, a solution that augments traditional DVFS processor frequency tuning with two additional workload-specific knobs: 1) speculative decoding and 2) variable verification depth to expand the optimization space to multiple dimensions for more power efficient on-device LLM inference. In extensive evaluations across hardware platforms and datasets, PELM demonstrates superior performance compared to state-of-the-art power governing methods, with up to 23.1% speedup and 52.4% reduction in energy consumption, while maintaining comparable task performance. The source code is available at https://github.com/imec-nu/PELM.

cs.LG↗

arXiv2026

Depth Exploration for LLM Decoding

Autoregressive LLM decoding evaluates every generated token through the full layer stack, even though many tokens become predictable at intermediate depths. Existing lossless depth-adaptive methods exploit this redundancy by choosing a single non-final exit depth and verifying its prediction with the final-depth model. However, our measurements show that this selection-based strategy leaves substantial headroom: choosing an exit too late wastes computation, while choosing one too early triggers fallback and discards dependent drafts. We propose Depth Exploration Decoding (DEX), a lossless decoding algorithm that replaces single-depth selection with parallel exploration over multiple candidate depths. At each commit position, DEX validates candidates against the final-depth reference, commits exactly the final-depth token, and collapses the exploration lattice to retain only reusable branch states. This expand--commit--collapse procedure preserves equivalence to standard autoregressive decoding while reducing the cost of committing each token. Across early-exit-trained and standard LLMs, DEX outperforms representative depth-selection baselines and achieves competitive end-to-end throughput against speculative and distributed decoding methods. Moreover, DEX improves as the explored depths become finer, showing that parallel depth exploration provides a scalable way to exploit the underused depth axis of LLM decoding.

cs.LG↗

arXiv2026

Attention Drift: What Autoregressive Speculative Decoding Models Learn

Speculative decoding accelerates LLM inference by drafting future tokens with a small model, but drafter models degrade sharply under template perturbation and long-context inputs. We identify a previously-unreported phenomenon we call \textbf{attention drift}: as the drafter generates successive tokens within a speculation chain, attention progressively moves from the prompt onto its own recently-generated tokens. We observe this across both \emph{EAGLE3} drafters and \emph{MTP heads}, suggesting drift is a property of drafter designs. We trace this to the un-normalized residual path between chain steps: the drafter's hidden state magnitude grows monotonically with chain depth, which exhibits dynamics consistent with additional pre-norm transformer layers stacked on the target rather than as a standalone autoregressive predictor. In order to limit the growth, we propose two architectural changes: Post-norm on the drafter hidden states and per-hidden-state RMSNorm after capturing target hidden states. Our interventions improve acceptance length over the current leading model, pre-norm EAGLE3, by up to $2\times$ under template perturbation, $1.18\times$ on long-context tasks, and $1.10\times$ on seven standard benchmarks spanning multi-turn chat, math, and coding. Our changes also allow shorter train-time-test depths to generalize over longer drafting sequences.

cs.LG↗

arXiv2026

Modality-Aware Zero-Shot Pruning and Sparse Attention for Efficient Multimodal Edge Inference

Edge devices increasingly run multimodal sensing pipelines that must remain accurate despite fluctuating power budgets and unpredictable sensor dropout. Existing pruning methods fail under these conditions: they generally require fine-tuning after compression, consuming over $10\times$ the deployment energy, and they assign static importance scores that are blind to which sensors are present. We present the SentryFuse framework, which addresses both challenges jointly through two key components. First, SentryGate learns modality-conditioned importance scores during training via first-order saliency supervision and then prunes attention heads and feed-forward channels at deployment without fine-tuning. Second, SentryAttend replaces dense self-attention, a key bottleneck in contemporary multimodal architectures, with sparse grouped-query attention, yielding a net 15% reduction in GFLOPs across three different multimodal architectures. Across three applications and multimodal backbones, SentryGate achieves a 12.7% average accuracy improvement over the strongest pruning baseline, and upto to 18% under modality dropout conditions. Together, SentryFuse reduces memory by 28.2% and lowers latency by up to $1.63\times$ without further fine-tuning, establishing modality-aware zero-shot compression as a practical path to multimodal intelligence on heterogeneous edge hardware.

cs.LG↗

arXiv2026

UNIFERENCE: A Discrete Event Simulation Framework for Developing Distributed AI Models

Developing and evaluating distributed inference algorithms remains difficult due to the lack of standardized tools for modeling heterogeneous devices and networks. Existing studies often rely on ad-hoc testbeds or proprietary infrastructure, making results hard to reproduce and limiting exploration of hypothetical hardware or network configurations. We present UNIFERENCE, a discrete-event simulation (DES) framework designed for developing, benchmarking, and deploying distributed AI models within a unified environment. UNIFERENCE models device and network behavior through lightweight logical processes that synchronize only on communication primitives, eliminating rollbacks while preserving the causal order. It integrates seamlessly with PyTorch Distributed, enabling the same codebase to transition from simulation to real deployment. Our evaluation demonstrates that UNIFERENCE profiles runtime with up to 98.6% accuracy compared to real physical deployments across diverse backends and hardware setups. By bridging simulation and deployment, UNIFERENCE provides an accessible, reproducible platform for studying distributed inference algorithms and exploring future system designs, from high-performance clusters to edge-scale devices. The framework is open-sourced at https://github.com/Dogacel/Uniference.

cs.DC↗

arXiv2025

MAESTRO : Adaptive Sparse Attention and Robust Learning for Multimodal Dynamic Time Series

From clinical healthcare to daily living, continuous sensor monitoring across multiple modalities has shown great promise for real-world intelligent decision-making but also faces various challenges. In this work, we introduce MAESTRO, a novel framework that overcomes key limitations of existing multimodal learning approaches: (1) reliance on a single primary modality for alignment, (2) pairwise modeling of modalities, and (3) assumption of complete modality observations. These limitations hinder the applicability of these approaches in real-world multimodal time-series settings, where primary modality priors are often unclear, the number of modalities can be large (making pairwise modeling impractical), and sensor failures often result in arbitrary missing observations. At its core, MAESTRO facilitates dynamic intra- and cross-modal interactions based on task relevance, and leverages symbolic tokenization and adaptive attention budgeting to construct long multimodal sequences, which are processed via sparse cross-modal attention. The resulting cross-modal tokens are routed through a sparse Mixture-of-Experts (MoE) mechanism, enabling black-box specialization under varying modality combinations. We evaluate MAESTRO against 10 baselines on four diverse datasets spanning three applications, and observe average relative improvements of 4% and 8% over the best existing multimodal and multivariate approaches, respectively, under complete observations. Under partial observations -- with up to 40% of missing modalities -- MAESTRO achieves an average 9% improvement. Further analysis also demonstrates the robustness and efficiency of MAESTRO's sparse, modality-aware design for learning from dynamic time series.

cs.LG↗

arXiv2025

MambaLite-Micro: Memory-Optimized Mamba Inference on MCUs

Deploying Mamba models on microcontrollers (MCUs) remains challenging due to limited memory, the lack of native operator support, and the absence of embedded-friendly toolchains. We present, to our knowledge, the first deployment of a Mamba-based neural architecture on a resource-constrained MCU, a fully C-based runtime-free inference engine: MambaLite-Micro. Our pipeline maps a trained PyTorch Mamba model to on-device execution by (1) exporting model weights into a lightweight format, and (2) implementing a handcrafted Mamba layer and supporting operators in C with operator fusion and memory layout optimization. MambaLite-Micro eliminates large intermediate tensors, reducing 83.0% peak memory, while maintaining an average numerical error of only 1.7x10-5 relative to the PyTorch Mamba implementation. When evaluated on keyword spotting(KWS) and human activity recognition (HAR) tasks, MambaLite-Micro achieved 100% consistency with the PyTorch baselines, fully preserving classification accuracy. We further validated portability by deploying on both ESP32S3 and STM32H7 microcontrollers, demonstrating consistent operation across heterogeneous embedded platforms and paving the way for bringing advanced sequence models like Mamba to real-world resource-constrained applications.

cs.LG↗

arXiv2025

Creativity or Brute Force? Using Brainteasers as a Window into the Problem-Solving Abilities of Large Language Models

Accuracy remains a standard metric for evaluating AI systems, but it offers limited insight into how models arrive at their solutions. In this work, we introduce a benchmark based on brainteasers written in long narrative form to probe more deeply into the types of reasoning strategies that models use. Brainteasers are well-suited for this goal because they can be solved with multiple approaches, such as a few-step solution that uses a creative insight or a longer solution that uses more brute force. We investigate large language models (LLMs) across multiple layers of reasoning, focusing not only on correctness but also on the quality and creativity of their solutions. We investigate many aspects of the reasoning process: (1) semantic parsing of the brainteasers into precise mathematical competition style formats; (2) generating solutions from these mathematical forms; (3) self-correcting solutions based on gold solutions; (4) producing step-by-step sketches of solutions; and (5) making use of hints. We find that LLMs are in many cases able to find creative, insightful solutions to brainteasers, suggesting that they capture some of the capacities needed to solve novel problems in creative ways. Nonetheless, there also remain situations where they rely on brute force despite the availability of more efficient, creative solutions, highlighting a potential direction for improvement in the reasoning abilities of LLMs.

cs.AI↗

arXiv2025

TW-CRL: Time-Weighted Contrastive Reward Learning for Efficient Inverse Reinforcement Learning

Episodic tasks in Reinforcement Learning (RL) often pose challenges due to sparse reward signals and high-dimensional state spaces, which hinder efficient learning. Additionally, these tasks often feature hidden "trap states" -- irreversible failures that prevent task completion but do not provide explicit negative rewards to guide agents away from repeated errors. To address these issues, we propose Time-Weighted Contrastive Reward Learning (TW-CRL), an Inverse Reinforcement Learning (IRL) framework that leverages both successful and failed demonstrations. By incorporating temporal information, TW-CRL learns a dense reward function that identifies critical states associated with success or failure. This approach not only enables agents to avoid trap states but also encourages meaningful exploration beyond simple imitation of expert trajectories. Empirical evaluations on navigation tasks and robotic manipulation benchmarks demonstrate that TW-CRL surpasses state-of-the-art methods, achieving improved efficiency and robustness.

cs.LG↗

arXiv2025

Towards On-Device Evidence Gathering for Intimate Partner Infiltration: A Feasibility Study for Joint Identity-Action Detection

Intimate Partner Infiltration (IPI) refers to phone-side privacy infiltration in intimate or close relationships, often enabled by physical access to a person's smartphone and discussed in technology-facilitated Intimate Partner Violence (IPV) contexts. Unlike conventional cyberattackers, IPI perpetrators leverage proximity and personal knowledge to circumvent standard protection, underscoring the need for targeted interventions, motivating device-side tools that surface such risk evidence for later review. While prior works have extensively studied IPV, and some have provided tailored and effective solutions such as security clinics, they are necessarily episodic and human-expert-intensive, and offer limited automated visibility into what happens on a smartphone between support sessions. Guided by a formative interview with experts (n=5), we take the first exploration into gathering IPI-risk evidence from a mobile system perspective and present AID, Automated IPI Detection, a data-driven system that continuously logs unauthorized access and suspicious behaviors on smartphones. In a controlled 27-participant study, AID achieves an end-to-end F1 score of 0.928 with a 7.0% false positive rate for Top-1 phone-side risk flagging; when preserving top-3 candidate action categories as report context, AID achieves an F1 score of 0.981 and a false positive rate of 1.6%. These findings demonstrate AID's potential as an evidence-support tool that complements current clinic-based interpretation and safety-planning.

cs.CR↗

arXiv2024

ElastiFormer: Learned Redundancy Reduction in Transformer via Self-Distillation

We introduce ElastiFormer, a post-training technique that adapts pretrained Transformer models into an elastic counterpart with variable inference time compute. ElastiFormer introduces small routing modules (as low as .00006% additional trainable parameters) to dynamically selects subsets of network parameters and input tokens to be processed by each layer of the pretrained network in an inputdependent manner. The routing modules are trained using self-distillation losses to minimize the differences between the output of the pretrained-model and their elastic counterparts. As ElastiFormer makes no assumption regarding the modality of the pretrained Transformer model, it can be readily applied to all modalities covering causal language modeling, image modeling as well as visual-language modeling tasks. We show that 20% to 50% compute saving could be achieved for different components of the transformer layer, which could be further reduced by adding very low rank LoRA weights (rank 1) trained via the same distillation objective. Finally, by comparing routing trained on different subsets of ImageNet, we show that ElastiFormer is robust against the training domain.

cs.LG↗

arXiv2024

AIRA: A Low-cost IR-based Approach Towards Autonomous Precision Drone Landing and NLOS Indoor Navigation

Automatic drone landing is an important step for achieving fully autonomous drones. Although there are many works that leverage GPS, video, wireless signals, and active acoustic sensing to perform precise landing, autonomous drone landing remains an unsolved challenge for palm-sized microdrones that may not be able to support the high computational requirements of vision, wireless, or active audio sensing. We propose AIRA, a low-cost infrared light-based platform that targets precise and efficient landing of low-resource microdrones. AIRA consists of an infrared light bulb at the landing station along with an energy efficient hardware photodiode (PD) sensing platform at the bottom of the drone. AIRA costs under 83 USD, while achieving comparable performance to existing vision-based methods at a fraction of the energy cost. AIRA requires only three PDs without any complex pattern recognition models to accurately land the drone, under $10$cm of error, from up to $11.1$ meters away, compared to camera-based methods that require recognizing complex markers using high resolution images with a range of only up to $1.2$ meters from the same height. Moreover, we demonstrate that AIRA can accurately guide drones in low light and partial non line of sight scenarios, which are difficult for traditional vision-based approaches.

cs.RO↗

arXiv2024

TRAMBA: A Hybrid Transformer and Mamba Architecture for Practical Audio and Bone Conduction Speech Super Resolution and Enhancement on Mobile and Wearable Platforms

We propose TRAMBA, a hybrid transformer and Mamba architecture for acoustic and bone conduction speech enhancement, suitable for mobile and wearable platforms. Bone conduction speech enhancement has been impractical to adopt in mobile and wearable platforms for several reasons: (i) data collection is labor-intensive, resulting in scarcity; (ii) there exists a performance gap between state of-art models with memory footprints of hundreds of MBs and methods better suited for resource-constrained systems. To adapt TRAMBA to vibration-based sensing modalities, we pre-train TRAMBA with audio speech datasets that are widely available. Then, users fine-tune with a small amount of bone conduction data. TRAMBA outperforms state-of-art GANs by up to 7.3% in PESQ and 1.8% in STOI, with an order of magnitude smaller memory footprint and an inference speed up of up to 465 times. We integrate TRAMBA into real systems and show that TRAMBA (i) improves battery life of wearables by up to 160% by requiring less data sampling and transmission; (ii) generates higher quality voice in noisy environments than over-the-air speech; (iii) requires a memory footprint of less than 20.0 MB.

cs.SD↗

arXiv2024

FlexiFly: Interfacing the Physical World with Foundation Models Empowered by Reconfigurable Drone Systems

Foundation models (FM) have shown immense human-like capabilities for generating digital media. However, foundation models that can freely sense, interact, and actuate the physical domain is far from being realized. This is due to 1) requiring dense deployments of sensors to fully cover and analyze large spaces, while 2) events often being localized to small areas, making it difficult for FMs to pinpoint relevant areas of interest relevant to the current task. We propose FlexiFly, a platform that enables FMs to ``zoom in'' and analyze relevant areas with higher granularity to better understand the physical environment and carry out tasks. FlexiFly accomplishes by introducing 1) a novel image segmentation technique that aids in identifying relevant locations and 2) a modular and reconfigurable sensing and actuation drone platform that FMs can actuate to ``zoom in'' with relevant sensors and actuators. We demonstrate through real smart home deployments that FlexiFly enables FMs and LLMs to complete diverse tasks up to $85\%$ more successfully. FlexiFly is critical step towards FMs and LLMs that can naturally interface with the physical world.

cs.RO↗

arXiv2021

Deep Learning in Human Activity Recognition with Wearable Sensors: A Review on Advances

Mobile and wearable devices have enabled numerous applications, including activity tracking, wellness monitoring, and human--computer interaction, that measure and improve our daily lives. Many of these applications are made possible by leveraging the rich collection of low-power sensors found in many mobile and wearable devices to perform human activity recognition (HAR). Recently, deep learning has greatly pushed the boundaries of HAR on mobile and wearable devices. This paper systematically categorizes and summarizes existing work that introduces deep learning methods for wearables-based HAR and provides a comprehensive analysis of the current advancements, developing trends, and major challenges. We also present cutting-edge frontiers and future directions for deep learning-based HAR.

cs.HC↗