arXiv subjects

Adam Lee

Publications and source records attributed to Adam Lee.

12 recordsLinked to original sources

arXiv2026

DexSIM: Real-time Dexterous Simulation with Unified Causal Video Diffusion

Recent progress of video diffusion models have enabled extensive simulation of the physical world. While simulation with hand object interaction has been less explored. We propose DexSIM, a dexterous simulation framework for simulating dexterous manipulation in real-time. While previous works utilizing video diffusion and 3D reconstruction focus on navigation, dexterous manipulation has been limited while it has extensive applications for creating interactive experiences with the simulated world and for generating synthetic data for robotics. Existing methods lack real-time interactivity and long-term spatial consistency and memory. We propose a 2-stage training framework for DexSIM. First we train a bi-directional video diffusion model by jointly embedding the hand action trajectory and video in a unified feature space. We utilize gaussian heatmap hand encoding for more accurate hand representation. Then we conduct a roll-out based autoregressive training with updated spatial cache as attention sink for spatial memory, which improves long-term consistency and 3D aware dexterous manipulation simulation. DexSIM outperforms the baseline on pixel and semantic similarity, motion fidelity, and hand projection accuracy. It also allows new applications such as hand motion transfer and runs at 15.24 FPS real-time interactivity.

cs.CV↗

arXiv2025

GOATex: Geometry & Occlusion-Aware Texturing

We present GOATex, a diffusion-based method for 3D mesh texturing that generates high-quality textures for both exterior and interior surfaces. While existing methods perform well on visible regions, they inherently lack mechanisms to handle occluded interiors, resulting in incomplete textures and visible seams. To address this, we introduce an occlusion-aware texturing framework based on the concept of hit levels, which quantify the relative depth of mesh faces via multi-view ray casting. This allows us to partition mesh faces into ordered visibility layers, from outermost to innermost. We then apply a two-stage visibility control strategy that progressively reveals interior regions with structural coherence, followed by texturing each layer using a pretrained diffusion model. To seamlessly merge textures obtained across layers, we propose a soft UV-space blending technique that weighs each texture's contribution based on view-dependent visibility confidence. Empirical results demonstrate that GOATex consistently outperforms existing methods, producing seamless, high-fidelity textures across both visible and occluded surfaces. Unlike prior works, GOATex operates entirely without costly fine-tuning of a pretrained diffusion model and allows separate prompting for exterior and interior mesh regions, enabling fine-grained control over layered appearances. For more qualitative results, please visit our project page: https://goatex3d.github.io/.

cs.CV↗

arXiv2025

MIMIC-MJX: Neuromechanical Emulation of Animal Behavior

The primary output of the nervous system is movement and behavior. While recent advances have democratized pose tracking during complex behavior, kinematic trajectories alone provide only indirect access to the underlying control processes. Here we present MIMIC-MJX, a framework for learning biomechanically grounded neural control policies from kinematics. MIMIC-MJX provides a platform for modeling the generative process of motor control by training neural controllers that learn to actuate biomechanical animal models in physics simulation to reproduce real kinematic trajectories. We demonstrate that our implementation is accurate, fast, and generalizable to diverse animal body models, and that it can be trained with modest amounts of motion data. MIMIC-MJX can be used to model motor control policies and simulate behavioral experiments, illustrating its potential as an integrative modeling framework for neuroscience.

q-bio.NC↗

arXiv2025

SIMSplat: Language-Aligned 4D Gaussian Splatting for Driving Scenario Generation

Driving scene manipulation using real-world sensor data has emerged as a promising alternative to traditional driving simulators. Despite advances in language control and neural scene representations, existing methods treat grounding, editing, and simulation as loosely connected stages, relying on heuristic object localization, manual guidance, and single-agent validation, thereby constraining semantic expressiveness and hindering scalable, reactive scenario generation. We introduce SIMSplat, a driving scene editor built on scene-graph-based 4D Gaussian Splatting augmented with language-aligned features. By embedding appearance, motion, and location semantics directly into Gaussian scene-graph nodes, SIMSplat makes reconstructed scenes queryable through free-form natural language, bridging language understanding to object-level editing and multi-agent simulation within a single framework. Building on this language-grounded scene graph, SIMSplat supports diverse edits including fine-grained pedestrian manipulation, while a multi-agent path refinement module propagates changes across all agents to ensure reactive, physically plausible simulations. The pipeline further integrates with Vision-Language Models for automated scenario mining. Experiments show that SIMSplat more than doubles baseline grounding accuracy, achieves the highest task completion rate, and produces the lowest failure rates across diverse driving scenarios.

cs.RO↗

arXiv2025

Semiparametrics via parametrics and contiguity

Inference on the parametric part of a semiparametric model is no trivial task. If one approximates the infinite dimensional part of the semiparametric model by a parametric function, one obtains a parametric model that is in some sense close to the semiparametric model and inference may proceed by the method of maximum likelihood. Under regularity conditions, the ensuing maximum likelihood estimator is asymptotically normal and efficient in the approximating parametric model. Thus one obtains a sequence of asymptotically normal and efficient estimators in a sequence of growing parametric models that approximate the semiparametric model and, intuitively, the limiting 'semiparametric' estimator should be asymptotically normal and efficient as well. In this paper we make this intuition rigorous: we move much of the semiparametric analysis back into classical parametric terrain, and then translate our parametric results back to the semiparametric world by way of contiguity. Our approach departs from the conventional sieve literature by being more specific about the approximating parametric models, by working not only with but also under these when treating the parametric models, and by taking full advantage of the mutual contiguity that we require between the parametric and semiparametric models. We illustrate our theory with two canonical examples of semiparametric models, namely the partially linear regression model and the Cox regression model. An upshot of our theory is a new, relatively simple, and rather parametric proof of the efficiency of the Cox partial likelihood estimator.

math.ST↗

arXiv2024

Towards Efficient Visual-Language Alignment of the Q-Former for Visual Reasoning Tasks

Recent advancements in large language models have demonstrated enhanced capabilities in visual reasoning tasks by employing additional encoders for aligning different modalities. While the Q-Former has been widely used as a general encoder for aligning several modalities including image, video, audio, and 3D with large language models, previous works on its efficient training and the analysis of its individual components have been limited. In this work, we investigate the effectiveness of parameter efficient fine-tuning (PEFT) the Q-Former using InstructBLIP with visual reasoning benchmarks ScienceQA and IconQA. We observe that applying PEFT to the Q-Former achieves comparable performance to full fine-tuning using under 2% of the trainable parameters. Additionally, we employ AdaLoRA for dynamic parameter budget reallocation to examine the relative importance of the Q-Former's sublayers with 4 different benchmarks. Our findings reveal that the self-attention layers are noticeably more important in perceptual visual-language reasoning tasks, and relative importance of FFN layers depends on the complexity of visual-language patterns involved in tasks. The code is available at https://github.com/AttentionX/InstructBLIP_PEFT.

cs.CL↗

arXiv2024

MERLIN: Multimodal Embedding Refinement via LLM-based Iterative Navigation for Text-Video Retrieval-Rerank Pipeline

The rapid expansion of multimedia content has made accurately retrieving relevant videos from large collections increasingly challenging. Recent advancements in text-video retrieval have focused on cross-modal interactions, large-scale foundation model training, and probabilistic modeling, yet often neglect the crucial user perspective, leading to discrepancies between user queries and the content retrieved. To address this, we introduce MERLIN (Multimodal Embedding Refinement via LLM-based Iterative Navigation), a novel, training-free pipeline that leverages Large Language Models (LLMs) for iterative feedback learning. MERLIN refines query embeddings from a user perspective, enhancing alignment between queries and video content through a dynamic question answering process. Experimental results on datasets like MSR-VTT, MSVD, and ActivityNet demonstrate that MERLIN substantially improves Recall@1, outperforming existing systems and confirming the benefits of integrating LLMs into multimodal retrieval systems for more responsive and context-aware multimedia retrieval.

cs.CL↗

arXiv2024

Locally Regular and Efficient Tests in Non-Regular Semiparametric Models

This paper considers hypothesis testing in semiparametric models which may be non-regular. I show that C($\alpha$) style tests are locally regular under mild conditions, including in cases where locally regular estimators do not exist, such as models which are (semiparametrically) weakly identified. I characterise the appropriate limit experiment in which to study local (asymptotic) optimality of tests in the non-regular case and generalise classical power bounds to this case. I give conditions under which these power bounds are attained by the proposed C($\alpha$) style tests. The application of the theory to a single index model and an instrumental variables model is worked out in detail.

econ.EM↗

arXiv2023

Reversible random number generation for adjoint Monte Carlo simulation of the heat equation

In PDE-constrained optimization, one aims to find design parameters that minimize some objective, subject to the satisfaction of a partial differential equation. A major challenges is computing gradients of the objective to the design parameters, as applying the chain rule requires computing the Jacobian of the design parameters to the PDE's state. The adjoint method avoids this Jacobian by computing partial derivatives of a Lagrangian. Evaluating these derivatives requires the solution of a second PDE with the adjoint differential operator to the constraint, resulting in a backwards-in-time simulation. Particle-based Monte Carlo solvers are often used to compute the solution to high-dimensional PDEs. However, such solvers have the drawback of introducing noise to the computed results, thus requiring stochastic optimization methods. To guarantee convergence in this setting, both the constraint and adjoint Monte Carlo simulations should simulate the same particle trajectories. For large simulations, storing full paths from the constraint equation for re-use in the adjoint equation becomes infeasible due to memory limitations. In this paper, we provide a reversible extension to the family of permuted congruential pseudorandom number generators (PCG). We then use such a generator to recompute these time-reversed paths for the heat equation, avoiding these memory issues.

math.NA↗

arXiv2022

Contrastive Weighted Learning for Near-Infrared Gaze Estimation

Appearance-based gaze estimation has been very successful with the use of deep learning. Many following works improved domain generalization for gaze estimation. However, even though there has been much progress in domain generalization for gaze estimation, most of the recent work have been focused on cross-dataset performance -- accounting for different distributions in illuminations, head pose, and lighting. Although improving gaze estimation in different distributions of RGB images is important, near-infrared image based gaze estimation is also critical for gaze estimation in dark settings. Also there are inherent limitations relying solely on supervised learning for regression tasks. This paper contributes to solving these problems and proposes GazeCWL, a novel framework for gaze estimation with near-infrared images using contrastive learning. This leverages adversarial attack techniques for data augmentation and a novel contrastive loss function specifically for regression tasks that effectively clusters the features of different samples in the latent space. Our model outperforms previous domain generalization models in infrared image based gaze estimation and outperforms the baseline by 45.6\% while improving the state-of-the-art by 8.6\%, we demonstrate the efficacy of our method.

cs.CV↗

arXiv2017

Leveraging Intel SGX to Create a Nondisclosure Cryptographic library

Enforcing integrity and confidentiality of users' application code and data is a challenging mission that any software developer working on an online production grade service is facing. Since cryptology is not a widely understood subject, people on the cutting edge of research and industry are always seeking for new technologies to naturally expand the security of their programs and systems. Intel Software Guard Extension (Intel SGX) is an Intel technology for developers who are looking to protect their software binaries from plausible attacks using hardware instructions. The Intel SGX puts sensitive code and data into CPU-hardened protected regions called enclaves. In this project we leverage the Intel SGX to produce a secure cryptographic library which keeps the generated keys inside an enclave restricting use and dissemination of confidential cryptographic keys. Using enclaves to store the keys we maintain a small Trusted Computing Base (TCB) where we also perform computation on temporary buffers to and from untrusted application code. As a proof of concept, we implemented hashes and symmetric encryption algorithms inside the enclave where we stored hashes, Initialization Vectors (IVs) and random keys and open sourced the code (https://github.com/hmofrad/CryptoEnclave).

cs.CR↗

arXiv2005

Defining a Comprehensive Threat Model for High Performance Computational Clusters

Over the past decade, high performance computational (HPC) clusters have become mainstream in academic and industrial settings as accessible means of computation. Throughout their proliferation, HPC security has been a secondary concern to performance. It is evident, however, that ensuring HPC security presents different challenges than the ones faced when dealing with traditional networks. To design suitable security measures for high performance computing, it is necessary to first realize the threats faced by such an environment. This task can be accomplished by the means of constructing a comprehensive threat model. To our knowledge, no such threat model exists with regards to Cluster Computing. In this paper, we explore the unique challenges of securing HPCs and propose a threat model based on the classical Confidentiality, Integrity and Availability security principles.

cs.CR↗