arXiv subjects

Chen Bai

Publications and source records attributed to Chen Bai.

At least 19 recordsLinked to original sources

arXiv2026

Data Scaling Laws in Stereo Matching: From In-domain Fitting to Cross-domain Generalization

In the era of large models, stereo matching is developing towards generalizable foundation models trained on million-scale synthetic images. However, how model performance scales up with synthetic data is still underexplored. In this work, we preliminarily investigate the in-domain and cross-domain data scaling laws in stereo matching. A realistic synthetic dataset, named XStereo, containing 200 thousand samples is first built to serve as a high-quality benchmark for data scaling research. In-domain training on five large-scale datasets demonstrate a two-regime scaling behavior with data, transitioning from the standard power law to saturation, where the critical transition point is approximately invariant under a fixed model. Cross-domain performance exhibits a substantially weaker and less predictable dependence on the source-domain data scale. The distribution and coverage diversity are more effective for diminishing domain gaps. More instructive suggestions are provided for high-efficiency data construction and training. The insightful findings are significant verification and extension of the existing scaling laws in both language and vision, paving the way to efficient and stronger stereo matching. The XStereo dataset is public at https://huggingface.co/datasets/Tong-Zhao/XStereo.

cs.CV↗

arXiv2026

FADiff: Fusion-Aware Differentiable Optimization for DNN Scheduling on Tensor Accelerators

Efficiently deploying Deep Neural Networks (DNNs) - particularly Large Language Models (LLMs) - on tensor accelerators is essential for achieving high throughput and energy efficiency at the edge. However, the combinatorial explosion of the joint design space spanning intra-layer mapping and inter-layer fusion renders existing approaches inadequate. In this work, we present FADiff, a gradient-based optimization framework that automatically discovers optimal mapping and fusion strategies to accelerate DNN inference. We first formulate a unified, differentiable analytical cost model that accurately estimates the energy and latency of individual layer mappings as well as diverse fusion configurations. We then combine CP-SAT-based pre-enumeration of feasible discrete candidates with augmented-Lagrangian penalties over continuous relaxations, enabling a constrained gradient-based search that efficiently navigates the vast design space to identify optimal joint strategies. Experimental results show that FADiff consistently outperforms prior methods, delivering superior optimization quality with significantly faster convergence in both energy and latency.

cs.AR↗

arXiv2026

TPCS: A Cycle-Level Simulation Paradigm with Native Execution Semantics for RTL Modeling and Generation

The growing diversity of domain-specific architectures and microarchitectural designs demands fast and accurate evaluation methods during early design stages. However, existing approaches suffer from a fundamental trade-off: high-level simulators offer speed but lack fidelity, while RTL-based tools provide accuracy at the cost of significant simulation overhead. This gap limits efficient pre-RTL microarchitecture exploration. Therefore, we propose Transaction-Pruned Cycle Simulation (TPCS), a novel modeling and simulation framework. It serves as a unified intermediate abstraction spanning microarchitecture modeling, cycle-accurate simulation, and synthesizable RTL implementation. TPCS organizes combinational logic as transaction-based function invocations and leverages host CPU control flow to naturally prune inactive logic, eliminating the need for explicit event scheduling or signal tracking. This enables cycle-accurate simulation with significantly improved performance. TPCS preserves structural correspondence with RTL and supports automatic generation of synthesizable designs. TPCS achieves 36$\times$-42$\times$ speedups over parallel Verilator on SMP systems and 3.4$\times$-7.2$\times$ over single-threaded Verilator on accelerators, with cycle-level waveform agreement on the evaluated traces. The generated RTL demonstrates comparable timing and area to handcrafted implementations, with relative changes of -4.7\% to +2.6\% in clock period and -3.3\% to +4.0\% in standard-cell area. By bridging the abstraction gap between high-level modeling and RTL design, TPCS enables a unified and agile pre-RTL workflow, allowing efficient microarchitecture exploration with both high performance and implementation fidelity.

cs.AR↗

arXiv2026

BenchmarkAnything: Agent-Driven Construction of Simulator-Ready Microarchitecture Benchmarks

The selection of benchmark workloads is of paramount importance in computer architecture, as it establishes the yardstick against which architectural innovations are measured and guided. Yet for decades, the SPEC benchmark suites, comprising merely tens of workloads, have been the de facto standard in academic architectural research, where they are frequently treated as a principal evaluation and optimization target. When a suite this small is relied upon so heavily, it risks architectural overfitting; as our research and prior studies demonstrate, an overly narrow focus can mislead design decisions by overvaluing certain innovations, producing cores that excel on SPEC benchmarks yet underperform on broader, realistic workloads. To mitigate this overfitting, adopting a large, comprehensive benchmark suite is the natural solution. However, the immense engineering effort required to strip software into the clean, interference-free binary executables demanded by simulators often makes this highly impractical. In this work, we demonstrate that AI agents provide an elegant solution to this challenge. Rather than manually curating yet another static benchmark suite, we introduce an agent-driven workflow capable of autonomously transforming arbitrary open-source repositories into simulator-ready executables. This automated approach makes workload collection highly scalable, allowing us to rapidly harvest hundreds of diverse applications from public repositories into our benchmark suite. Through a comparative analysis of our agent-generated suite against SPEC, we show that it not only achieves higher-fidelity performance assessments but also uncovers novel architectural insights that traditional, static suites fail to expose.

cs.AR↗

arXiv2026

F4R: Failure-Driven Recognition, Reconstruction, Refinement, and Redeployment for Continual Robot Self-Improvement

The real-world performance of current vision-language-action models is fundamentally constrained by the limited coverage of expert demonstrations and their insufficient understanding of physical interactions. A common remedy is to collect additional real-world demonstrations of newly encountered failures. However, this process is costly, inefficient, potentially unsafe, and difficult to scale. To address this challenge, we propose Failure for Rising (F4R), a failure-driven real-to-sim-to-real closed-loop learning framework that converts real-world failures into targeted policy improvement. F4R first uses an agent to automatically identify and diagnose failures from rollouts. It reconstructs each failure as an interactive, object-centric table-top environment that preserves the task-relevant spatial and physical conditions. The policy is then refined through failure-conditioned sim-real co-training followed by targeted reinforcement learning in the reconstructed environments. The improved policy is subsequently redeployed, while newly observed failures are continuously fed back into the next reconstruction and learning cycle. Real-world evaluations on four manipulation tasks show that F4R achieves 93.75% In-Distribution and 90.0% Out-of-Distribution (OOD) success, outperforming the budget-matched Targeted BC baseline by 18.75 percentage points under OOD conditions without collecting additional real-world corrective demonstrations.

cs.RO↗

arXiv2026

Bootstrapping Symmetries in Quantum Many-Body Systems from the Cross Spectral Form Factor

Symmetries play a central role in quantum many-body physics, yet uncovering them systematically remains challenging. We introduce a bootstrap framework to reconstruct the representation theory of hidden finite group symmetries of quantum many-body lattice Hamiltonians, using only a known symmetry subgroup $N$ and spectral correlations between its sectors. We introduce the cross spectral form factor (xSFF) as a subgroup-resolved diagnostic of hidden symmetry, computed via exact diagonalization to seed the bootstrap. Applying constraints from these data alongside the algebraic conditions of the fusion rules, our bootstrap sharply restricts candidate groups $G$. Without prior assumptions about $G$, our method can systematically recover its representation-theoretic data, including the number and dimensions of irreducible representations, their branching rules with respect to $N$, the fusion algebra, and the full character table. This framework applies to chaotic and integrable systems and accommodates unitary and anti-unitary symmetries. Through examples, we demonstrate that $G$ can be uniquely identified. Our bootstrap recovers the $\mathbb{Z}_4$ symmetry at the self-dual point of the three-state quantum torus chain, detects signatures of projective representations in the effective Hamiltonian of the driven Bose-Hubbard model, and rediscovers the $η$-pairing $\mathrm{SO}(4)$ symmetry of the one-dimensional Fermi-Hubbard model. Leveraging the universal random-matrix ramp in chaotic systems, we demonstrate the possibility of ramp-only bootstrap alongside an example with hidden commuting symmetries. We establish the xSFF as a diagnostic for conserved order parameters in a symmetry-breaking spin-$1$ chain and introduce an ancilla-assisted measurement circuit for experimental realization. Our framework establishes a practical route to identify symmetries from dynamical spectral observables.

quant-ph↗

arXiv2026

Extracting central charge from ground-state overlaps of spatially deformed Hamiltonians

We show that the conformal anomaly of a $(1+1)$-dimensional conformal field theory can be extracted directly from a ground-state wave-function overlap associated with a spatial conformal deformation. Focusing on the $q$-Möbius deformation, we derive an exact overlap formula between the deformed and undeformed ground states, whose exponent directly encodes the central charge. Motivated by this result, we construct a lattice estimator based solely on ground-state overlaps and apply it to representative critical quantum chains and the gapless edge modes of a two-dimensional Chern insulator. Numerical results demonstrate that the resulting overlaps provide a simple and robust probe of the central charge in microscopic models. We further demonstrate that the deformed ground states retain universal geometric structures in their entanglement spectra and entanglement entropies. These results provide a simple wave-function-based route to probing conformal data in critical systems and topological edge modes.

cond-mat.str-el↗

arXiv2026

MicroEvo: Knowledge-Guided LLM Sampling for Efficient Microarchitecture Design Space Exploration

Microarchitecture design space exploration suffers from expansive search spaces and expensive PPA evaluation, leaving only a small simulation budget for design decision-making. Existing methods perform blind search without considering microarchitectural dependencies and fail to learn from the iterative search effectively, leading to wasted evaluations and weak Pareto convergence. In this paper, we propose MicroEvo, a knowledge-guided framework that couples off-the-shelf LLMs with Monte Carlo Tree Search (MCTS) for multi-objective microarchitecture optimization. MicroEvo combines LLM-driven evolutionary operators, a Pareto-aware tree policy that balances Pareto contribution and diversity, an active knowledge accumulation mechanism that extracts and reuses optimization insights, and state-aware directives that adapt the search behavior online. Experiments show that MicroEvo improves Pareto-front quality by up to 36.2% over NSGA-II and achieves 10.6x higher search efficiency, and also demonstrates strong scalability to a complex industrial-scale core. The code repository is available at: https://github.com/GEAR-SEU/MicroEvo-ICCAD-26.

cs.AI↗

arXiv2026

Bridging the Information Gap: Semantic Densification and Hindsight Distillation for Cold-Start Prediction

New-user cold-start is a critical bottleneck for e-commerce platforms: predicting user lifetime value (LTV) and conversion rate (CVR) for users with sparse interaction history. Two prior directions -- LLM-based semantic augmentation and learning using privileged information (LUPI) -- each face a key limitation. First, LLM augmentation produces unstructured rationales that are noisy and hard to operationalize in production. Second, naive student-teacher distillation can be brittle due to an information gap between the privileged teacher and the sparse student; moreover, this gap is heterogeneous across users. We propose SemRaD, a Semantic Reasoning-aware Distillation framework addressing both limitations. First, a Structured Semantic Reasoning Pipeline replaces free-form rationales with a structured schema built via a discover-curate-audit workflow, producing per user a Densified Semantic Profile (consumed by the deployed student via a Semantic-Gated Encoder that focuses on the most informative dimensions) and a Hindsight Distillation Target reconciled from pre- and post-conversion reasoning (used only at training). Second, to bridge this gap and handle its heterogeneity, a Hindsight-Aware Distillation Network transfers privileged knowledge via the hindsight target, with Distillation Experts improving transfer under per-user variability. On a large-scale industrial dataset, SemRaD lifts +1.9% LTV (Gini) and +1.0% CVR (AUROC) over a production-grade base; a four-week online A/B at Keeta confirms +1.0% LTV / +0.43% CVR. SemRaD also matches the production system's LTV using only 9% of the training data while improving CVR by 0.8%.

cs.AI↗

arXiv2026

Exact Lindbladian Dynamics from Conformal Embeddings and Topological Defects in Conformal Field Theory

Analyzing the dynamics of physical observables in open quantum many-body systems is a fundamental but highly challenging task that has yielded very few exact results. In this work, we identify intrinsic conformal structures that restore exact solvability in $(1+1)$D conformal field theories. For $N$ Majorana fermions with linear mode jumps, the adjoint Lindbladian is triangular on reduced even Majorana monomials, yielding recursive exact Heisenberg evolution. In Wess-Zumino-Witten models admitting conformal Majorana embeddings, this hierarchy gives exact dynamics of affine-current products realized as Majorana bilinears, including regimes where the Kac-Moody current algebra alone does not close. In diagonal rational conformal field theories, Verlinde topological defect lines furnish jump operators whose primary-sector dynamics is exactly diagonal: topological-charge probabilities are conserved, while intersector coherences dephase at rates fixed by the modular $S$ matrix and nonnegative measurement strengths. These examples show that intrinsic conformal structures, such as conformal embeddings and modular data, can organize exactly solvable open conformal dynamics.

cond-mat.stat-mech↗

arXiv2026

Drive-JEPA: Video JEPA Meets Multimodal Trajectory Distillation for End-to-End Driving

End-to-end autonomous driving increasingly leverages self-supervised video pretraining to learn transferable planning representations. However, pretraining video world models for scene understanding has so far brought only limited improvements. This limitation is compounded by the inherent ambiguity of driving: each scene typically provides only a single human trajectory, making it difficult to learn multimodal behaviors. In this work, we propose Drive-JEPA, a framework that integrates Video Joint-Embedding Predictive Architecture (V-JEPA) with multimodal trajectory distillation for end-to-end driving. First, we adapt V-JEPA for end-to-end driving, pretraining a ViT encoder on large-scale driving videos to produce predictive representations aligned with trajectory planning. Second, we introduce a proposal-centric planner that distills diverse simulator-generated trajectories alongside human trajectories, with a momentum-aware selection mechanism to promote stable and safe behavior. When evaluated on NAVSIM, the V-JEPA representation combined with a simple transformer-based decoder outperforms prior methods by 3 PDMS in the perception-free setting. The complete Drive-JEPA framework achieves 93.3 PDMS on v1 and 87.8 EPDMS on v2, setting a new state-of-the-art.

cs.CV↗

arXiv2026

From Uncertainty to Stability and Fidelity: Guiding Sparse-View 3D Gaussian Splatting with Fisher Information

3D Gaussian Splatting (3DGS) has emerged as a promising technique for novel view synthesis. However, 3DGS requires dense input views to achieve high-quality rendering. In sparse-view scenarios, 3DGS often prones to overfitting, resulting in noticeable artifacts and degraded rendering quality. Previous methods explore to address this issue by introducing additional priors (e.g. depth priors) or integrating regularization techniques (e.g. Dropout). However, these methods are often applied without principled guidance. In particular, prior-based augmentation typically samples novel viewpoints randomly, while Dropout-based regularization randomly removes Gaussians. The compounded randomness introduces uncertainty and instability, limiting the fidelity of novel view synthesis. In this paper, we propose a novel method for sparse-view 3DGS that incorporates Fisher Information to quantitatively guide the utilization of geometric priors and regularization. Specifically, our method comprises two key components: (1) Stereo augmentation with Fisher Information. By leveraging Fisher Information, we actively select most informative supporting views and use depth priors to curate reliable pseudo ground truths, which reduces randomness in augmentation and improves stability and rendering fidelity; (2) Uncertainty-aware regularization. We reduce the instability of Dropout-based regularization by using Fisher Information to quantitatively measure the uncertainty of each 3D Gaussian, and adaptively adjust the removal probability, leading to more stable and effective regularization. With these two components, our method effectively mitigates overfitting and improves the stability of optimization in sparse-view 3DGS, resulting in superior rendering fidelity. Extensive experiments show that our method achieves state-of-the-art performance in sparse-view novel view synthesis benchmarks.

cs.CV↗

arXiv2026

CPPL: A Circuit Prompt Programming Language

Large language models (LLMs) have shown promise in register-transfer level (RTL) design automation, but direct RTL generation remains difficult to validate, optimize, and integrate with compiler-based hardware design flows. Hardware compiler infrastructures such as CIRCT provide typed intermediate representations, legality checks, and optimization passes, yet current LLMs struggle to emit raw compiler IR because of MLIR syntax, SSA discipline, dialect-specific operations, and strict width constraints. This paper presents CPPL, a compiler-mediated design framework that turns LLM-assisted hardware generation into a statically checkable frontend problem rather than an unconstrained RTL text-generation task. CPPL combines a Python frontend DSL for declaring module interfaces and hierarchy with CPPL IR, a JSON-based circuit IR designed to expose compiler-visible structure while remaining accessible to LLMs. The compiler infers operation widths from declared module ports, validates generated IR, checks hierarchy and port bindings, and deterministically lowers the result to CIRCT for synthesizable Verilog generation. On the RTLLM benchmark, CPPL improves functional correctness over direct Verilog and direct CIRCT IR generation, while CIRCT optimization reduces post-synthesis AIG node counts. These results show that a compiler-mediated interface can make LLM-assisted hardware design more reliable, analyzable, and amenable to backend optimization. CPPL is available at https://github.com/SawyDust1228/CPPL.

cs.AR↗

arXiv2026

PipeRTL: Timing-Aware Pipeline Optimization at IR-Level for RTL Generation

Modern hardware compilers increasingly rely on rich intermediate representations (IRs) to preserve optimization-relevant semantics before generating RTL code. However, one important optimization is still largely deferred to backend tools: pipeline optimization. In common RTL flows, registers are inserted by frontend heuristics or hardware designers and later adjusted by backend retiming after the design has been lowered to a much lower-level netlist representation. At that point, much of the operator-level structure originally exposed by the compiler IR has already been weakened or lost, limiting opportunities for global, compiler-level pipeline optimization. This paper presents PipeRTL, an IR-level pipeline optimization framework for hardware compilers, instantiated in CIRCT. PipeRTL makes the legality of register relocation explicit in the IR, uses a learned timing predictor to approximate downstream delay behavior, and formulates timing-aware register relocation as a global min-cost flow problem under timing constraints. Evaluation on open-source designs under a commercial backend synthesis flow shows that PipeRTL improves downstream implementation quality on average, reducing critical-path delay, power, and area across the evaluated benchmarks, while also providing a stronger starting point for backend retiming. These results indicate that exposing pipeline optimization as an explicit compiler pass can deliver backend-meaningful gains by improving the sequential structure presented to later stages and the resulting downstream implementation quality.

cs.AR↗

arXiv2026

Post-Quantum Security of Block Cipher Constructions

Block ciphers are versatile cryptographic ingredients that are used in a wide range of applications ranging from secure Internet communications to disk encryption. While post-quantum security of public-key cryptography has received significant attention, the case of symmetric-key cryptography (and block ciphers in particular) remains a largely unexplored topic. In this work, we set the foundations for a theory of post-quantum security for block ciphers and associated constructions. Leveraging our new techniques, we provide the first post-quantum security proofs for the key-length extension scheme FX, the tweakable block ciphers LRW and XEX, and most block cipher encryption and authentication modes. Our techniques can be used for security proofs in both the plain model and the quantum ideal cipher model. Our work takes significant initial steps in establishing a rigorous understanding of the post-quantum security of practical symmetric-key cryptography.

cs.CR↗

arXiv2026

DrivePTS: A Progressive Learning Framework with Textual and Structural Enhancement for Driving Scene Generation

Synthesis of diverse driving scenes serves as a crucial data augmentation technique for validating the robustness and generalizability of autonomous driving systems. Current methods aggregate high-definition (HD) maps and 3D bounding boxes as geometric conditions in diffusion models for conditional scene generation. However, implicit inter-condition dependency causes generation failures when control conditions change independently. Additionally, these methods suffer from insufficient details in both semantic and structural aspects. Specifically, brief and view-invariant captions restrict semantic contexts, resulting in weak background modeling. Meanwhile, the standard denoising loss with uniform spatial weighting neglects foreground structural details, causing visual distortions and blurriness. To address these challenges, we propose DrivePTS, which incorporates three key innovations. Firstly, our framework adopts a progressive learning strategy to mitigate inter-dependency between geometric conditions, reinforced by an explicit mutual information constraint. Secondly, a Vision-Language Model is utilized to generate multi-view hierarchical descriptions across six semantic aspects, providing fine-grained textual guidance. Thirdly, a frequency-guided structure loss is introduced to strengthen the model's sensitivity to high-frequency elements, improving foreground structural fidelity. Extensive experiments demonstrate that our DrivePTS achieves state-of-the-art fidelity and controllability in generating diverse driving scenes. Notably, DrivePTS successfully generates rare scenes where prior methods fail, highlighting its strong generalization ability.

cs.CV↗

arXiv2026

Relaxation Process During Complex Time Evolution In Two-Dimensional Integrable and Chaotic CFTs

We investigate the complex time evolution of a vacuum state with the insertion of a local primary operator in two-dimensional conformal field theories (2d CFTs). This complex time evolution can be considered as a composite process constructed from Lorentzian time evolution and a Euclidean evolution induced by a post-selected measurement. Our main finding is that in the spatially-compact system, this complex time evolution drives the state of the subsystems to those of the primary state with the same conformal dimensions of the inserted operator. Contrary to the compact system, the subsystems of the spatially non-compact system evolve to states that depend on the non-unitary process during a certain time regime. In holographic systems with a compact spatial direction, this process induced by a heavy local operator can correspond to the relaxation from a black hole with an inhomogeneous horizon to that with a uniform one, while in the ones with a non-compact spatial direction, it can correspond to the relaxation to that with a horizon depending on the non-unitary process.

hep-th↗

arXiv2025

Spatially Structured Entanglement from Nonequilibrium Thermal Pure States

We study quantum quench dynamics in (1+1)-dimensional critical systems, starting from thermal pure states called crosscap states, and evolving them under spatially inhomogeneous Hamiltonians. The spatial inhomogeneity is introduced through a deformation of the Hamiltonian, expressed as linear combinations of the generators of the $SL^{(q)}(2,\mathbb{R})$ subalgebra of the Virasoro algebra. We analyze the free massless Dirac fermion theory and holographic conformal field theory as prototypical examples of integrable and non-integrable dynamics. Consistent with general expectations, "Möbius-type" deformations lead to thermalization in the non-integrable case, and to periodic revivals in the integrable one. In contrast, "sine-square-type" and "displacement-type" deformations prevent both thermalization and scrambling, instead producing late-time, graph-like entanglement patterns. These patterns emerge from the interplay between the deformed Hamiltonian and the crosscap initial state and appear to be universal: they are determined solely by the deformation profile while remaining largely insensitive to microscopic details. Finally, we perform a holographic calculation in three-dimensional gravity using AdS$_3$/CFT$_2$, which reproduces the main features of our (1+1)-dimensional study.

cond-mat.stat-mech↗