arXiv subjects

Yichen Liu

Publications and source records attributed to Yichen Liu.

At least 55 records · Page 3Linked to original sources

arXiv2026

Unusual Dual Flat Bands and two-dimensional Dirac-node Arc State in Kagome Metal Ni3In2S2

Kagome materials are at the frontier of condensed matter physics. An ideal kagome lattice features only one geometrically frustrated flat band spanning the entire momentum space and a single Dirac cone at the Brillouin-zone corners. However, for the first time, here we observe unusual flat-band and Dirac physics in the newly discovered "322" kagome material Ni3In2S2 by combining high-resolution synchrotron- and laser-based angle-resolved photoemission spectroscopy with a micro-focused beam, scanning tunneling microscopy, and first-principles calculations. We resolve two distinct electronic flat-band states located in close proximity to the Fermi level: a robust Topological Surface Flat Band at ~40 meV below the Fermi level on the Sulfur-terminated surface, originating from weak topological insulator states, and a kagome lattice-derived flat band at ~100 meV binding energy with an ultranarrow bandwidth (~5 meV). Instead of the single Dirac cone, the Indium-terminated surface hosts a rare two-dimensional Dirac-node arc state, where the gapless Dirac nodes extend along an open one-dimensional line crossing the Brillouin-zone boundary, exhibiting sharp linear dispersion, exceptionally high Fermi velocity, and pronounced circular dichroism. These findings establish Ni3In2S2 as a unique topological kagome metal in which multiple flat-band states of different physical origin coexist with an unusual Dirac-node arc, opening an avenue for discovering flat-band--driven and topology-enabled quantum phenomena.

cond-mat.mtrl-sci↗

arXiv2026

Shedding the envelope: JWST reveals a kiloparsec-scale [OIII]-weak Balmer shell around a z=7.64 quasar

Luminous quasars at the redshift frontier z>7 serve as stringent probes of super-massive black hole formation and they are thought to undergo much of their growth obscured by dense gas and dust in their host galaxies. Fully characterizing the symbiotic evolution of SMBHs and hosts requires rest-frame optical observations that span spatial scales from the broad-line region to the ISM and CGM. JWST now provides the necessary spatially resolved spectroscopy to do so. But the physical conditions that regulate the interplay between SMBHs and their hosts at the highest redshifts, especially the nature of early feedback phases, remain unclear. We present JWST/NIRSpec IFU observations of J0313$-$1806 at z=7.64, the most distant luminous quasar known. From the restframe optical spectrum of the unresolved quasar, we derive a black hole mass of $M_\mathrm{BH}=(1.63 \pm 0.10)\times10^9 M_\odot$ based on H$β$ and an Eddington rate of $λ=L/L_\mathrm{Edd}=0.80\pm 0.05$, consistent with previous MgII-based estimates. J0313-1806 exhibits no detectable [O III] emission on nuclear scales. Most remarkably, we detect an ionized gas shell extending out to $\sim 1.8$ kpc traced by H$β$ emission that also lacks any significant [O III], with a $3σ$ upper limit on the [O III]$ λ$5007 to H$β$ flux ratio of $\log_{10} \left( F(\mathrm{[OIII]})/F(\mathrm{H}β)\right)=-1.15$. Through photoionization modelling, we demonstrate that the extended emission is consistent with a thin, clumpy outflowing shell where [OIII] is collisionally de-excited by dense gas. We interpret this structure as a fossil remnant of a recent blowout phase, providing evidence for episodic feedback cycles in one of the earliest quasars. These findings suggest that dense ISM phases may play a crucial role in shaping the spectral properties of quasars accross cosmic time.

astro-ph.GA↗

arXiv2025

Control of Microrobots with Reinforcement Learning under On-Device Compute Constraints

An important function of autonomous microrobots is the ability to perform robust movement over terrain. This paper explores an edge ML approach to microrobot locomotion, allowing for on-device, lower latency control under compute, memory, and power constraints. This paper explores the locomotion of a sub-centimeter quadrupedal microrobot via reinforcement learning (RL) and deploys the resulting controller on an ultra-small system-on-chip (SoC), SC$μ$M-3C, featuring an ARM Cortex-M0 microcontroller running at 5 MHz. We train a compact FP32 multilayer perceptron (MLP) policy with two hidden layers ($[128, 64]$) in a massively parallel GPU simulation and enhance robustness by utilizing domain randomization over simulation parameters. We then study integer (Int8) quantization (per-tensor and per-feature) to allow for higher inference update rates on our resource-limited hardware, and we connect hardware power budgets to achievable update frequency via a cycles-per-update model for inference on our Cortex-M0. We propose a resource-aware gait scheduling viewpoint: given a device power budget, we can select the gait mode (trot/intermediate/gallop) that maximizes expected RL reward at a corresponding feasible update frequency. Finally, we deploy our MLP policy on a real-world large-scale robot on uneven terrain, qualitatively noting that domain-randomized training can improve out-of-distribution stability. We do not claim real-world large-robot empirical zero-shot transfer in this work.

cs.RO↗

arXiv2025

The Discovery of Little Red Dots in the Local Universe: Signatures of Cool Gas Envelopes

JWST observations have revealed a population of high-redshift "little red dots" (LRDs) that challenge conventional AGN models. We report the discovery of three local LRDs at $z = 0.1$-$0.2$, initially selected from the SDSS database, with follow-up optical/near-IR spectroscopy and photometry. They exhibit properties fully consistent with those of high-redshift LRDs, including broad hydrogen and helium emission lines, compact morphologies, V-shaped UV-optical SED, declining near-IR continua, and no significant variability. Two sources were targeted but not detected in X-rays with statistical significance. All three sources show blue-shifted He I absorption, while two exhibit H$α$ and Na D absorption lines. We detect full Balmer and Paschen line series in all three objects, along with abundant narrow [Fe II] emission in two. The emission line analyses suggest narrow lines originate from AGN-powered, metal-poor regions with minimal dust; broad lines come from inner regions with exceptionally high density or atypical dust properties; and [Fe II] emission arises from dense gas between broad and narrow-line regions. One of our objects, J1025+1402 (nicknamed $The~Egg$), shows extremely high equivalent width Na D, K I, and Ca II triplet absorption lines, along with other potential low-ionization absorption features, suggesting the presence of a cool ($\sim$5000 K), metal-enriched gas envelope. The optical/near-IR continua of these LRDs are also consistent with theoretical models featuring an atmosphere around black holes. The WISE-detected IR emission is consistent with weak dust emission of $T \sim 10^2-10^3$ K. We propose a conceptual model consisting of a largely thermalized cool-gas envelope surrounding the central black hole and an extended emission line region with high-density outflowing gas to explain the observed properties of these local LRDs.

astro-ph.GA↗

arXiv2025

Bridging Quasars and Little Red Dots: Insights into Broad-Line AGNs at $z=5-8$ from the First JWST COSMOS-3D Dataset

We report the discovery of 13 broad-line AGNs at $z = 5 - 8$ from the first 10% data of the JWST Cycle 3 Treasury Program COSMOS-3D. These AGNs are identified by their broad H$α$ or H$β$ emission lines through the NIRCam grism wide-field slitless spectroscopy. One object at $z = 7.646$ with broad H$β$ emission has an F444W magnitude of 23.6 mag, making it one of the brightest $z > 7.5$ broad-line AGNs yet known. Among the 13 AGNs, 10 objects have reddened optical continua with slopes $β_{\rm opt}>0$. The remaining three objects have their overall SEDs that resemble those of UV-luminous quasars at similar redshifts, but their $β_{\rm opt}$, though negative, are not as blue as those of unobscured quasars. We also obtain MIRI photometry at 7.7-18 $μ$m for two AGNs and place strong constraints on their rest-frame near-IR SED. We find no significant variability in the rest-frame UV by comparing the COSMOS-3D and COSMOS-Web F115W images taken apart by 60 days in the rest-frame. We compute the H$α$ luminosity functions (LFs) for the broad H$α$ emitters at $z \approx 5-6$ and find a potential redshift evolution when compared with that of the $z \approx 4-5$ sample. We also derive the H$β$ LF at $z\sim8$ for AGNs and galaxies by combining our sample with those from the literature. The broad H$β$ emitters in this work suggest a number density two orders of magnitude higher than that predicted by the quasar LF based on rest-frame UV-selected samples. As a preview, our work showcases the ability of the COSMOS-3D grism survey to provide a complete view of the properties, growth, and evolution of bright broad-line AGNs at $z>5$.

astro-ph.GA↗

arXiv2025

Evaluation of Large Language Models for Numeric Anomaly Detection in Power Systems

Large language models (LLMs) have gained increasing attention in power grids for their general-purpose capabilities. Meanwhile, anomaly detection (AD) remains critical for grid resilience, requiring accurate and interpretable decisions based on multivariate telemetry. Yet the performance of LLMs on large-scale numeric data for AD remains largely unexplored. This paper presents a comprehensive evaluation of LLMs for numeric AD in power systems. We use GPT-OSS-20B as a representative model and evaluate it on the IEEE 14-bus system. A standardized prompt framework is applied across zero-shot, few-shot, in-context learning, low rank adaptation (LoRA), fine-tuning, and a hybrid LLM-traditional approach. We adopt a rule-aware design based on the three-sigma criterion, and report detection performance and rationale quality. This study lays the groundwork for further investigation into the limitations and capabilities of LLM-based AD and its integration with classical detectors in cyber-physical power grid applications.

eess.SY↗

arXiv2025

Language-Instructed Reasoning for Group Activity Detection via Multimodal Large Language Model

Group activity detection (GAD) aims to simultaneously identify group members and categorize their collective activities within video sequences. Existing deep learning-based methods develop specialized architectures (e.g., transformer networks) to model the dynamics of individual roles and semantic dependencies between individuals and groups. However, they rely solely on implicit pattern recognition from visual features and struggle with contextual reasoning and explainability. In this work, we propose LIR-GAD, a novel framework of language-instructed reasoning for GAD via Multimodal Large Language Model (MLLM). Our approach expand the original vocabulary of MLLM by introducing an activity-level token and multiple cluster-specific tokens. We process video frames alongside two specially designed tokens and language instructions, which are then integrated into the MLLM. The pretrained commonsense knowledge embedded in the MLLM enables the token and tokens to effectively capture the semantic information of collective activities and learn distinct representational features of different groups, respectively. Also, we introduce a multi-label classification loss to further enhance the token's ability to learn discriminative semantic representations. Then, we design a Multimodal Dual-Alignment Fusion (MDAF) module that integrates MLLM's hidden embeddings corresponding to the designed tokens with visual features, significantly enhancing the performance of GAD. Both quantitative and qualitative experiments demonstrate the superior performance of our proposed method in GAD taks.

cs.CV↗

arXiv2025

A Deep Chandra X-ray Survey of a Luminous Quasar Sample at $z\sim$ 7

We present new Chandra observations of seven luminous quasars at $z>6.5$. Combined with archival Chandra observations of all other known quasars, they form nearly complete X-ray observations of all currently known $z\sim7$ quasars with $M_{1450}<-26.5$, except for J0313$-$1806 at $z=7.642$ and J0910$-$0414 at $z=6.636$. Together with existing ground-based NIR spectroscopy and ALMA observations, we investigate the correlations between X-ray emission (the X-ray luminosity $L_{\rm X}$ and the optical/UV-to-X-ray spectral slope $α_{\rm OX}$) and various quasar properties (rest-UV luminosity $L_{\mathrm{2500\ \mathring{A}}}$, bolometric luminosity $L_{\rm bol}$, C IV blueshift, and infrared luminosity $L_{\rm IR}$). We find most $z>6.5$ quasars follow a similar $α_{\rm OX}-L_{\mathrm{2500\ \mathring{A}}}$ relation as $z\sim1-6$ quasars, but also display a large scatter. We find a potential correlation between $α_{\rm OX}$ and the C IV blueshift, suggesting a soft optical/UV-to-X-ray SED shape is frequently associated with fast disk winds. Furthermore, we analyze the X-ray spectrum of 11 quasars at $z>6.5$ with Chandra detection, and find the best-fit photon index $Γ$ is $2.41\pm0.27$, which is likely driven by high accretion rates of $z>6.5$ quasars. In addition, we find there are no significant correlations between either $L_{\rm X}$ and $L_{\rm IR}$, nor $L_{\rm bol}$ and $L_{\rm IR}$, suggesting no strong correlations between quasar luminosity and star formation luminosity for the most luminous quasars at $z>6.5$.

astro-ph.HE↗

arXiv2025

Unconventional Magneto-Optical Effects in Altermagnets

The ideal altermagnets are a class of collinear, crystal-symmetry-enforced fully compensated magnets with nonrelativistic spin-split bands, in which contributions from Berry curvature to magneto-optical effects (MOEs) are strictly forbidden by an effective time-reversal symmetry. Here we show that, in such systems, MOEs are exclusively induced by the quantum metric and, in realistic altermagnets, are typically dominated by it. We refer to Berry-curvature-induced MOEs as conventional MOEs and to quantum-metric-dominated MOEs as unconventional MOEs. We derive general formulas that incorporate both Berry curvature and quantum metric for unconventional MOEs in altermagnets, enabling a quantitative evaluation of their respective contributions. Through symmetry analysis, we prove that ideal altermagnets are constrained to exhibit only unconventional MOEs. Using the three-dimensional canonical altermagnet MnTe and the emerging two-dimensional bilayer twisted altermagnet CrSBr as illustrative examples, we demonstrate that unconventional MOEs are prevalent in altermagnets. Our results establish altermagnets as a natural platform for quantum-metric-driven optical phenomena, substantially broadening the scope of MOEs and providing concrete predictions that can be tested in future experimental studies.

cond-mat.mtrl-sci↗

arXiv2025

TrajMamba: An Efficient and Semantic-rich Vehicle Trajectory Pre-training Model

Vehicle GPS trajectories record how vehicles move over time, storing valuable travel semantics, including movement patterns and travel purposes. Learning travel semantics effectively and efficiently is crucial for real-world applications of trajectory data, which is hindered by two major challenges. First, travel purposes are tied to the functions of the roads and points-of-interest (POIs) involved in a trip. Such information is encoded in textual addresses and descriptions and introduces heavy computational burden to modeling. Second, real-world trajectories often contain redundant points, which harm both computational efficiency and trajectory embedding quality. To address these challenges, we propose TrajMamba, a novel approach for efficient and semantically rich vehicle trajectory learning. TrajMamba introduces a Traj-Mamba Encoder that captures movement patterns by jointly modeling both GPS and road perspectives of trajectories, enabling robust representations of continuous travel behaviors. It also incorporates a Travel Purpose-aware Pre-training procedure to integrate travel purposes into the learned embeddings without introducing extra overhead to embedding calculation. To reduce redundancy in trajectories, TrajMamba features a Knowledge Distillation Pre-training scheme to identify key trajectory points through a learnable mask generator and obtain effective compressed trajectory embeddings. Extensive experiments on two real-world datasets and three downstream tasks show that TrajMamba outperforms state-of-the-art baselines in both efficiency and accuracy.

cs.LG↗

arXiv2025

CVD-STORM: Cross-View Video Diffusion with Spatial-Temporal Reconstruction Model for Autonomous Driving

Generative models have been widely applied to world modeling for environment simulation and future state prediction. With advancements in autonomous driving, there is a growing demand not only for high-fidelity video generation under various controls, but also for producing diverse and meaningful information such as depth estimation. To address this, we propose CVD-STORM, a cross-view video diffusion model utilizing a spatial-temporal reconstruction Variational Autoencoder (VAE) that generates long-term, multi-view videos with 4D reconstruction capabilities under various control inputs. Our approach first fine-tunes the VAE with an auxiliary 4D reconstruction task, enhancing its ability to encode 3D structures and temporal dynamics. Subsequently, we integrate this VAE into the video diffusion process to significantly improve generation quality. Experimental results demonstrate that our model achieves substantial improvements in both FID and FVD metrics. Additionally, the jointly-trained Gaussian Splatting Decoder effectively reconstructs dynamic scenes, providing valuable geometric information for comprehensive scene understanding. Our project page is https://sensetime-fvg.github.io/CVD-STORM.

cs.CV↗

arXiv2025

The enhanced X-ray Timing and Polarimetry mission -- eXTP for launch in 2030

In this paper we present the current status of the enhanced X-ray Timing and Polarimetry mission, which has been fully approved for launch in 2030. eXTP is a space science mission designed to study fundamental physics under extreme conditions of matter density, gravity, and magnetism. The mission aims at determining the equation of state of matter at supra-nuclear density, measuring the effects of quantum electro-dynamics, and understanding the dynamics of matter in strong-field gravity. In addition to investigating fundamental physics, the eXTP mission is poised to become a leading observatory for time-domain and multi-messenger astronomy in the 2030's, as well as providing observations of unprecedented quality on a variety of galactic and extragalactic objects. After briefly introducing the history and a summary of the scientific objectives of the eXTP mission, this paper presents a comprehensive overview of: 1) the cutting-edge technology, technical specifications, and anticipated performance of the mission's scientific instruments; 2) the full mission profile, encompassing spacecraft design, operational capabilities, and ground segment infrastructure.

astro-ph.HE↗

arXiv2025

First constraints on the local ionization topology in front of two quasars at z ~ 7.5

Thus far, Lyman-$α$ damping wings towards quasars have been used to probe the \textit{global} ionization state of the foreground intergalactic medium (IGM). A new parameterization has demonstrated that the damping wing signature also carries \textit{local} information about the distribution of neutral hydrogen (HI) in front of the quasar before it started shining. Leveraging a recently introduced Bayesian \texttt{JAX}-based Hamiltonian Monte Carlo (HMC) inference framework, we derive constraints on the Lorentzian-weighted HI column density $N_\mathrm{HI}^\mathrm{DW}$, the quasar's distance $r_\mathrm{patch}$ to the first neutral patch and its lifetime $t_\mathrm{Q}$ based on JWST/NIRSpec spectra of the two $z \sim 7.5$ quasars J1007+2115 and J1342+0928. After folding in model-dependent topology information, we find that J1007+2115 (and J1342+0928) is most likely to reside in a $\langle x_\mathrm{HI} \rangle = 0.32_{-0.20}^{+0.22}$ ($0.58_{-0.23}^{+0.23}$) neutral IGM while shining for a remarkably short lifetime of $\log_{10} t_\mathrm{Q} /\mathrm{yr} = 4.14_{-0.18}^{+0.74}$ (an intermediate lifetime of $5.64_{-0.43}^{+0.25}$) along a sightline with $\log_{10} N_\mathrm{HI}^\mathrm{DW} /\mathrm{cm}^{-2} = 19.70_{-0.86}^{+0.35}$ ($20.24_{-0.22}^{+0.25}$) and $r_\mathrm{patch} = 28.9_{-14.4}^{+54.0} \,\mathrm{cMpc}$ ($10.9_{-5.9}^{+5.6} \,\mathrm{cMpc}$). In light of the potential presence of local absorbers in the foreground of J1342+0928 as has been recently suggested, we also demonstrate how the Lorentzian-weighted column density $N_\mathrm{HI}^\mathrm{DW}$ provides a natural means for quantifying their contribution to the observed damping wing signal.

astro-ph.CO↗

arXiv2025

The Effects of Communication Delay on Human Performance and Neurocognitive Responses in Mobile Robot Teleoperation

Communication delays in mobile robot teleoperation adversely affect human-machine collaboration. Understanding delay effects on human operational performance and neurocognition is essential for resolving this issue. However, no previous research has explored this. To fill this gap, we conduct a human-in-the-loop experiment involving 10 participants, integrating electroencephalography (EEG) and robot behavior data under varying delays (0-500 ms in 100 ms increments) to systematically investigate these effects. Behavior analysis reveals significant performance degradation at 200-300 ms delays, affecting both task efficiency and accuracy. EEG analysis discovers features with significant delay dependence: frontal $θ/β$-band and parietal $α$-band power. We also identify a threshold window (100-200 ms) for early perception of delay in humans, during which these EEG features first exhibit significant differences. When delay exceeds 400 ms, all features plateau, indicating saturation of cognitive resource allocation at physiological limits. These findings provide the first evidence of perceptual and cognitive delay thresholds during teleoperation tasks in humans, offering critical neurocognitive insights for the design of delay compensation strategies.

cs.RO↗

arXiv2025

LIFBench: Evaluating the Instruction Following Performance and Stability of Large Language Models in Long-Context Scenarios

As Large Language Models (LLMs) evolve in natural language processing (NLP), their ability to stably follow instructions in long-context inputs has become critical for real-world applications. However, existing benchmarks seldom focus on instruction-following in long-context scenarios or stability on different inputs. To bridge this gap, we introduce LIFBench, a scalable dataset designed to evaluate LLMs' instruction-following capabilities and stability across long contexts. LIFBench comprises three long-context scenarios and eleven diverse tasks, featuring 2,766 instructions generated through an automated expansion method across three dimensions: length, expression, and variables. For evaluation, we propose LIFEval, a rubric-based assessment method that enables precise, automated scoring of complex LLM responses without reliance on LLM-assisted assessments or human judgment. This method allows for a comprehensive analysis of model performance and stability from multiple perspectives. We conduct detailed experiments on 20 prominent LLMs across six length intervals. Our work contributes LIFBench and LIFEval as robust tools for assessing LLM performance in complex and long-context settings, offering valuable insights to guide future advancements in LLM development.

cs.CL↗

arXiv2025

ManuSearch: Democratizing Deep Search in Large Language Models with a Transparent and Open Multi-Agent Framework

Recent advances in web-augmented large language models (LLMs) have exhibited strong performance in complex reasoning tasks, yet these capabilities are mostly locked in proprietary systems with opaque architectures. In this work, we propose \textbf{ManuSearch}, a transparent and modular multi-agent framework designed to democratize deep search for LLMs. ManuSearch decomposes the search and reasoning process into three collaborative agents: (1) a solution planning agent that iteratively formulates sub-queries, (2) an Internet search agent that retrieves relevant documents via real-time web search, and (3) a structured webpage reading agent that extracts key evidence from raw web content. To rigorously evaluate deep reasoning abilities, we introduce \textbf{ORION}, a challenging benchmark focused on open-web reasoning over long-tail entities, covering both English and Chinese. Experimental results show that ManuSearch substantially outperforms prior open-source baselines and even surpasses leading closed-source systems. Our work paves the way for reproducible, extensible research in open deep search systems. We release the data and code in https://github.com/RUCAIBox/ManuSearch

cs.CL↗

arXiv2025

SAPPHIRES: Extremely Metal-Poor Galaxy Candidates with $12+{\rm log(O/H)}<7.0$ at $z\sim5-7$ from Deep JWST/NIRCam Grism Observations

Population III stars, the hypothetical first generation metal-free stars, have yet to be discovered. Even after three years of successful JWST operations, studies have shown that most galaxies identified to date at $z > 5$ exhibit a metallicity floor of $Z\gtrsim2\%\,Z_{\odot}$, possibly due to unknown selection biases toward bright galaxies or rapid metal enrichment. To address this question, we search for galaxies with low R3$=$[OIII]$\,λ$5008/H$β$ ratios as part of the JWST Cycle-3 large treasury program, the Slitless Areal Pure-Parallel HIgh-Redshift Emission Survey (SAPPHIRES). Using deep NIRCam Wide-Field Slitless Spectrscopy (WFSS) data, we report the discovery of seven extremely metal-poor galaxy candidates in the SAPPHIRES Early Data Release field, with estimated $12+{\rm log(O/H)}<7.0$ at $z\sim5-7$, including two sources with $Z<1\%\,Z_{\odot}$, significantly breaking the metallicity floor observed both locally and at high redshift. These candidates appear extremely faint ($\sim28-30\,$ F200W AB mag) and low-mass (${\rm log}(M_{*}/M_{\odot})\sim6.8-7.8$), as expected from the mass-metallicity relation. They also exhibit very blue UV slopes ($-2.6\lesssimβ\lesssim-2.0$), likely due to low dust content $A_{V}\lesssim0.2\,{\rm mag}$ or young stellar ages $\sim5-20\,{\rm Myr}$. Compared to galaxies at similar redshift, they appear exceptionally bursty in their star formation activity. Our results highlight the power of NIRCam/WFSS in identifying extremely metal-poor galaxies, from just a single pointing, with more data to come in SAPPHIRES. This underscores the potential of pure-parallel programs towards achieving JWST's primary science goal: discovering the first pristine stars and galaxies. Deep JWST/NIRSpec follow-up observations will also be essential to confirm their nature and perform detailed chemical abundance analyses.

astro-ph.GA↗

arXiv2025

Dynamic Regularized CBDT: Variance-Calibrated Causal Boosting for Interpretable Heterogeneous Treatment Effects

Heterogeneous treatment effect estimation in high-stakes applications demands models that simultaneously optimize precision, interpretability, and calibration. Many existing tree-based causal inference techniques, however, exhibit high estimation errors when applied to observational data because they struggle to capture complex interactions among factors and rely on static regularization schemes. In this work, we propose Dynamic Regularized Causal Boosted Decision Trees (CBDT), a novel framework that integrates variance regularization and average treatment effect calibration into the loss function of gradient boosted decision trees. Our approach dynamically updates the regularization parameters using gradient statistics to better balance the bias-variance tradeoff. Extensive experiments on standard benchmark datasets and real-world clinical data demonstrate that the proposed method significantly improves estimation accuracy while maintaining reliable coverage of true treatment effects. In an intensive care unit patient triage study, the method successfully identified clinically actionable rules and achieved high accuracy in treatment effect estimation. The results validate that dynamic regularization can effectively tighten error bounds and enhance both predictive performance and model interpretability.

cs.LG↗