arXiv subjects

Hongwei Ma

Publications and source records attributed to Hongwei Ma.

6 recordsLinked to original sources

arXiv2025

From Noise to Laws: Regularized Time-Series Forecasting via Denoised Dynamic Graphs

Long-horizon multivariate time-series forecasting is challenging because realistic predictions must (i) denoise heterogeneous signals, (ii) track time-varying cross-series dependencies, and (iii) remain stable and physically plausible over long rollout horizons. We present PRISM, which couples a score-based diffusion preconditioner with a dynamic, correlation-thresholded graph encoder and a forecast head regularized by generic physics penalties. We prove contraction of the induced horizon dynamics under mild conditions and derive Lipschitz bounds for graph blocks, explaining the model's robustness. On six standard benchmarks , PRISM achieves consistent SOTA with strong MSE and MAE gains.

cs.LG↗

arXiv2025

Signals, Concepts, and Laws: Toward Universal, Explainable Time-Series Forecasting

Accurate, explainable and physically credible forecasting remains a persistent challenge for multivariate time-series whose statistical properties vary across domains. We propose DORIC, a Domain-Universal, ODE-Regularized, Interpretable-Concept Transformer for Time-Series Forecasting that generates predictions through five self-supervised, domain-agnostic concepts while enforcing differentiable residuals grounded in first-principles constraints.

cs.LG↗

arXiv2025

PREIG: Physics-informed and Reinforcement-driven Interpretable GRU for Commodity Demand Forecasting

Accurately forecasting commodity demand remains a critical challenge due to volatile market dynamics, nonlinear dependencies, and the need for economically consistent predictions. This paper introduces PREIG, a novel deep learning framework tailored for commodity demand forecasting. The model uniquely integrates a Gated Recurrent Unit (GRU) architecture with physics-informed neural network (PINN) principles by embedding a domain-specific economic constraint: the negative elasticity between price and demand. This constraint is enforced through a customized loss function that penalizes violations of the physical rule, ensuring that model predictions remain interpretable and aligned with economic theory. To further enhance predictive performance and stability, PREIG incorporates a hybrid optimization strategy that couples NAdam and L-BFGS with Population-Based Training (POP). Experiments across multiple commodities datasets demonstrate that PREIG significantly outperforms traditional econometric models (ARIMA,GARCH) and deep learning baselines (BPNN,RNN) in both RMSE and MAPE. When compared with GRU,PREIG maintains good explainability while still performing well in prediction. By bridging domain knowledge, optimization theory and deep learning, PREIG provides a robust, interpretable, and scalable solution for high-dimensional nonlinear time series forecasting in economy.

cs.LG↗

arXiv2022

Lexical semantics enhanced neural word embeddings

Current breakthroughs in natural language processing have benefited dramatically from neural language models, through which distributional semantics can leverage neural data representations to facilitate downstream applications. Since neural embeddings use context prediction on word co-occurrences to yield dense vectors, they are inevitably prone to capture more semantic association than semantic similarity. To improve vector space models in deriving semantic similarity, we post-process neural word embeddings through deep metric learning, through which we can inject lexical-semantic relations, including syn/antonymy and hypo/hypernymy, into a distributional space. We introduce hierarchy-fitting, a novel semantic specialization approach to modelling semantic similarity nuances inherently stored in the IS-A hierarchies. Hierarchy-fitting attains state-of-the-art results on the common- and rare-word benchmark datasets for deriving semantic similarity from neural word embeddings. It also incorporates an asymmetric distance function to specialize hypernymy's directionality explicitly, through which it significantly improves vanilla embeddings in multiple evaluation tasks of detecting hypernymy and directionality without negative impacts on semantic similarity judgement. The results demonstrate the efficacy of hierarchy-fitting in specializing neural embeddings with semantic relations in late fusion, potentially expanding its applicability to aggregating heterogeneous data and various knowledge resources for learning multimodal semantic spaces.

cs.CL↗

arXiv2015

Triplet-Polaron Interaction Induced Upconversion from Triplet to Singlet: a New Way to Obtain Highly Efficient OLEDs

The triplet harvesting is a main challenge in organic light-emitting devices (OLEDs), due to the radiative decay of triplet is spin-forbidden. Here, we designed and synthesized two D-A type molecules, TPA-TAZ and TCP. The OLEDs based on them exhibit deep-blue emission and the singlet formation ratios are higher than the simple spin-statistics of 25 %. Specially, a TPA-TAZ-based OLED achieves a maximum EQE of 6.8 %, which is the largest value of the undoped OLEDs with CIE(y)< 0.06 (the EBU blue standard) up to date. Comprehensive experiments eliminate the triplet-harvesting processes of thermally activated delayed fluorescence and triplet-triplet annihilation. Instead, the triplet-polaron interaction induced upconversion from triplet to singlet through one-electron transfer mechanism is proposed, and proven by the magneto-current measurement and quantum chemistry computation. Our results may offer a new route to break through the 25 % upper limit of IQE of fluorescent OLEDs, especially, the deep-blue fluorescent OLEDs.

physics.chem-ph↗

arXiv2014

Organic light-emitting diodes using open-shell molecule as emitter: the emission from doublet

We fabricate OLEDs using a stable neutral π radical, BDPA, as the emitter. There is only one electron in the singly occupied molecular orbital (SOMO) of this open-shell molecule. This feature makes the excited state of open-shell molecules be neither singlet nor triplet, but doublet. The key issue of how to harvest the triplet energy in an OLED is thus bypassed, due to the radiative decay of doublet is totally spin allowed. In the BDPA-based OLED, the emission was confirmed to be from the electronic transition from LUMO to SOMO, via the frontier molecular orbital analysis combined with the spectroscopy measurements. The maximum luminance of the OLEDs is 4879 cd/m2 which is comparable to the first reported Fluorescence-, Phosphorecence- and TADF-based OLEDs.

physics.chem-ph↗