arXiv subjects

Michael Wu

Publications and source records attributed to Michael Wu.

12 recordsLinked to original sources

arXiv2026

Invalidation Contracts for Cross-Episode Agent Memory

LLM agents that cache recovery suggestions from API errors can skip re-derivation in later episodes, spending fewer tokens and fewer model calls on constraints they have already learned. Server-side data drift turns those cached fixes into silent failures, and the usual remedy, re-deriving on every episode, gives the savings back. We introduce invalidation contracts, a protocol layer that attaches version stamps and cacheability hints to every recovery suggestion so the client can evict stale entries without trial and error, and keep the rest. The contract decomposes realized savings into two independent factors: validity, the fraction of cached suggestions that remain correct after a drift event, and compliance, the fraction the planner applies on the first attempt. Validity depends only on the protocol and is vendor-independent. Compliance depends on the planner model: identical wire bytes yield 100% first-try compliance on Claude Haiku 4.5 and 11% or below on Claude Sonnet 5, which exhibits input-schema conservatism, refusing fixes that add fields the original request did not contain. We evaluate across seven models, three serving paths, two domains, and approximately 9,400 episodes. Row-level invalidation raises compliance by 0 to 66.7 percentage points across the seven models, 55.6 to 66.7 on three, and recovers 29-33% of baseline token cost on four of seven models, while table-level invalidation destroys co-located entries and drops post-drift first-try rates to 0% on five of seven. Eviction precision is 1.00 at row granularity on every model under the row-level oracle of Section 4.1. The contract adds 15% to response payload. Version-stamp validity is deterministic by construction and produced identical results across every model and serving path, with zero contract failures in the entire evaluation.

cs.AI↗

arXiv2026

FinCom: A Financial Multi-Agent Demo with Disagree-or-Commit Deliberation

Multi-agent systems powered by large language models (LLMs) are increasingly used for financial analysis and decision support. However, existing coordination schemes, especially those emphasizing consensus or debate, are vulnerable to sycophancy: agents conform to peer reasoning instead of evidence, leading to premature agreement and degraded outcomes. We introduce FinCom (Financial Committee), a governed multi-agent framework and interactive system that operationalizes the Disagree-or-Commit (DoC) protocol to embed structured dissent into financial AI committees. A central Supervisor orchestrates three ReAct-enabled specialist agents: Research, Quantitative, and Risk. Each agent is equipped with role-specific tools for retrieval, computation, and stress testing. During deliberation, agents must either explicitly critique or commit to their peers' reasoning before converging on a unified recommendation. This demonstration showcases how FinCom supports committee-style financial analysis through coordinated multi-agent interaction, including structured report generation and interactive decision support. Evaluated across the most recent financial agent benchmark, in addition to 90 internal handcrafted financial tasks using an LLM-as-a-Judge protocol, DoC improves reasoning accuracy and risk awareness significantly over a consensus-seeking baseline on both an in-house and external evaluation set. By reframing disagreement as a governance primitive rather than noise, FinCom offers a lightweight, prompt-only recipe for improving accountability, transparency, and epistemic robustness in agentic financial systems.

cs.MA↗

arXiv2026

TreeDGS: Aerial Gaussian Splatting for Distant DBH Measurement

Aerial remote sensing efficiently surveys large areas, but accurate direct object-level measurement remains difficult in complex natural scenes. Advancements in 3D computer vision, particularly radiance field representations such as NeRF and 3D Gaussian splatting, can improve reconstruction fidelity from posed imagery. Nevertheless, direct aerial measurement of important attributes like tree diameter at breast height (DBH) remains challenging. Trunks in aerial forest scans are distant and sparsely observed in image views; at typical operating altitudes, stems may span only a few pixels. With these constraints, conventional reconstruction methods have inaccurate breast-height trunk geometry. TreeDGS is an aerial image reconstruction method that uses 3D Gaussian splatting as a continuous scene representation for trunk measurement. After SfM--MVS initialization and Gaussian optimization, we extract a dense point set from the Gaussian field using RaDe-GS's depth-aware cumulative-opacity integration and associate each sample with a multi-view opacity reliability score. Then, we isolate trunk points and estimate DBH using opacity-weighted solid-circle fitting. Evaluated on 10 plots with field-measured DBH, TreeDGS reaches 4.79 cm RMSE (about 2.6 pixels at this GSD) and outperforms a LiDAR baseline (7.66 cm RMSE). This shows that TreeDGS can enable accurate, low-cost aerial DBH measurement .

cs.CV↗

arXiv2025

StruM: Structured Mixed Precision for Efficient Deep Learning Hardware Codesign

In this paper, we propose StruM, a novel structured mixed-precision-based deep learning inference method, co-designed with its associated hardware accelerator (DPU), to address the escalating computational and memory demands of deep learning workloads in data centers and edge applications. Diverging from traditional approaches, our method avoids time-consuming re-training/fine-tuning and specialized hardware access. By leveraging the variance in weight magnitudes within layers, we quantize values within blocks to two different levels, achieving up to a 50% reduction in precision for 8-bit integer weights to 4-bit values across various Convolutional Neural Networks (CNNs) with negligible loss in inference accuracy. To demonstrate efficiency gains by utilizing mixed precision, we implement StruM on top of our in-house FlexNN DNN accelerator [1] that supports low and mixed-precision execution. Experimental results depict that the proposed StruM-based hardware architecture achieves a 31-34% reduction in processing element (PE) power consumption and a 10% reduction in area at the accelerator level. In addition, the statically configured StruM results in 23-26% area reduction at the PE level and 2-3% area savings at the DPU level.

cs.AR↗

arXiv2021

OPTIMADE, an API for exchanging materials data

The Open Databases Integration for Materials Design (OPTIMADE) consortium has designed a universal application programming interface (API) to make materials databases accessible and interoperable. We outline the first stable release of the specification, v1.0, which is already supported by many leading databases and several software packages. We illustrate the advantages of the OPTIMADE API through worked examples on each of the public materials databases that support the full API specification.

cond-mat.mtrl-sci↗

arXiv2020

Trained Quantization Thresholds for Accurate and Efficient Fixed-Point Inference of Deep Neural Networks

We propose a method of training quantization thresholds (TQT) for uniform symmetric quantizers using standard backpropagation and gradient descent. Contrary to prior work, we show that a careful analysis of the straight-through estimator for threshold gradients allows for a natural range-precision trade-off leading to better optima. Our quantizers are constrained to use power-of-2 scale-factors and per-tensor scaling of weights and activations to make it amenable for hardware implementations. We present analytical support for the general robustness of our methods and empirically validate them on various CNNs for ImageNet classification. We are able to achieve near-floating-point accuracy on traditionally difficult networks such as MobileNets with less than 5 epochs of quantized (8-bit) retraining. Finally, we present Graffitist, a framework that enables automatic quantization of TensorFlow graphs for TQT (available at https://github.com/Xilinx/graffitist ).

cs.CV↗

arXiv2019

Affordance Analysis of Virtual and Augmented Reality Mediated Communication

Virtual and augmented reality communication platforms are seen as promising modalities for next-generation remote face-to-face interactions. Our study attempts to explore non-verbal communication features in relation to their conversation context for virtual and augmented reality mediated communication settings. We perform a series of user experiments, triggering nine conversation tasks in 4 settings, each containing corresponding non-verbal communication features. Our results indicate that conversation types which involve less emotional engagement are more likely to be acceptable in virtual reality and augmented reality settings with low-fidelity avatar representation, compared to scenarios that involve high emotional engagement or intellectually difficult discussions. We further systematically analyze and rank the impact of low-fidelity representation of micro-expressions, body scale, head pose, and hand gesture in affecting the user experience in one-on-one conversations, and validate that preserving micro-expression cues plays the most effective role in improving bi-directional conversations in future virtual and augmented reality settings.

cs.HC↗

arXiv2018

Quantizing Convolutional Neural Networks for Low-Power High-Throughput Inference Engines

Deep learning as a means to inferencing has proliferated thanks to its versatility and ability to approach or exceed human-level accuracy. These computational models have seemingly insatiable appetites for computational resources not only while training, but also when deployed at scales ranging from data centers all the way down to embedded devices. As such, increasing consideration is being made to maximize the computational efficiency given limited hardware and energy resources and, as a result, inferencing with reduced precision has emerged as a viable alternative to the IEEE 754 Standard for Floating-Point Arithmetic. We propose a quantization scheme that allows inferencing to be carried out using arithmetic that is fundamentally more efficient when compared to even half-precision floating-point. Our quantization procedure is significant in that we determine our quantization scheme parameters by calibrating against its reference floating-point model using a single inference batch rather than (re)training and achieve end-to-end post quantization accuracies comparable to the reference model.

cs.LG↗

arXiv2016

High-Throughput Data Detection for Massive MU-MIMO-OFDM using Coordinate Descent

Data detection in massive multi-user (MU) multiple-input multiple-output (MIMO) wireless systems is among the most critical tasks due to the excessively high implementation complexity. In this paper, we propose a novel, equalization-based soft-output data-detection algorithm and corresponding reference FPGA designs for wideband massive MU-MIMO systems that use orthogonal frequency-division multiplexing (OFDM). Our data-detection algorithm performs approximate minimum mean-square error (MMSE) or box-constrained equalization using coordinate descent. We deploy a variety of algorithm-level optimizations that enable near-optimal error-rate performance at low implementation complexity, even for systems with hundreds of base-station (BS) antennas and thousands of subcarriers. We design a parallel VLSI architecture that uses pipeline interleaving and can be parametrized at design time to support various antenna configurations. We develop reference FPGA designs for massive MU-MIMO-OFDM systems and provide an extensive comparison to existing designs in terms of implementation complexity, throughput, and error-rate performance. For a 128 BS antenna, 8 user massive MU-MIMO-OFDM system, our FPGA design outperforms the next-best implementation by more than 2.6x in terms of throughput per FPGA look-up tables.

cs.IT↗

arXiv2015

Linear Large-Scale MIMO Data Detection for 5G Multi-Carrier Waveform Candidates

Fifth generation (5G) wireless systems are expected to combine emerging transmission technologies, such as large-scale multiple-input multiple-output (MIMO) and non-orthogonal multi-carrier waveforms, to improve the spectral efficiency and to reduce out-of-band (OOB) emissions. This paper investigates the efficacy of two promising multi-carrier waveforms that reduce OOB emissions in combination with large-scale MIMO, namely filter bank multi-carrier (FBMC) and generalized frequency division multiplexing (GFDM). We develop novel, low-complexity data detection algorithms for both of these waveforms. We investigate the associated performance/complexity trade-offs in the context of large-scale MIMO, and we study the peak-to-average power ratio (PAPR). Our results show that reducing the OOB emissions with FBMC and GFDM leads to higher computational complexity and PAPR compared to that of orthogonal frequency-division multiplexing (OFDM) and single-carrier frequency division multiple access (SC-FDMA).

cs.IT↗

arXiv2014

Conjugate Gradient-based Soft-Output Detection and Precoding in Massive MIMO Systems

Massive multiple-input multiple-output (MIMO) promises improved spectral efficiency, coverage, and range, compared to conventional (small-scale) MIMO wireless systems. Unfortunately, these benefits come at the cost of significantly increased computational complexity, especially for systems with realistic antenna configurations. To reduce the complexity of data detection (in the uplink) and precoding (in the downlink) in massive MIMO systems, we propose to use conjugate gradient (CG) methods. While precoding using CG is rather straightforward, soft-output minimum mean-square error (MMSE) detection requires the computation of the post-equalization signal-to-interference-and-noise-ratio (SINR). To enable CG for soft-output detection, we propose a novel way of computing the SINR directly within the CG algorithm at low complexity. We investigate the performance/complexity trade-offs associated with CG-based soft-output detection and precoding, and we compare it to exact and approximate methods. Our results reveal that the proposed method outperforms existing algorithms for massive MIMO systems with realistic antenna configurations.

cs.IT↗

arXiv2014

Large-Scale MIMO Detection for 3GPP LTE: Algorithms and FPGA Implementations

Large-scale (or massive) multiple-input multiple-output (MIMO) is expected to be one of the key technologies in next-generation multi-user cellular systems, based on the upcoming 3GPP LTE Release 12 standard, for example. In this work, we propose - to the best of our knowledge - the first VLSI design enabling high-throughput data detection in single-carrier frequency-division multiple access (SC-FDMA)-based large-scale MIMO systems. We propose a new approximate matrix inversion algorithm relying on a Neumann series expansion, which substantially reduces the complexity of linear data detection. We analyze the associated error, and we compare its performance and complexity to those of an exact linear detector. We present corresponding VLSI architectures, which perform exact and approximate soft-output detection for large-scale MIMO systems with various antenna/user configurations. Reference implementation results for a Xilinx Virtex-7 XC7VX980T FPGA show that our designs are able to achieve more than 600 Mb/s for a 128 antenna, 8 user 3GPP LTE-based large-scale MIMO system. We finally provide a performance/complexity trade-off comparison using the presented FPGA designs, which reveals that the detector circuit of choice is determined by the ratio between BS antennas and users, as well as the desired error-rate performance.

cs.IT↗