arXiv subjects

Justin Wang

Publications and source records attributed to Justin Wang.

17 recordsLinked to original sources

arXiv2026

MARCUS: Missing-Aware Region Representation with Contextual Urban Signals for Rent Prediction

Multimodal urban data has expanded the applications of urban region representation learning, such as functional zone identification and real estate appraisal, but also introduces challenges caused by data incompleteness. Existing studies usually handle missing data through imputation, treating missingness as noise while ignoring its potential semantic value. To address this issue, we propose MARCUS, a missing-aware region representation model that treats missingness as a contextual urban signal. MARCUS models missingness in three stages: Intra Learning jointly encodes observed features and missing patterns, Inter Learning estimates modality reliability to guide cross-modal interaction, and Fusion uses missing-aware and time-aware gating to generate the final region embedding. We apply MARCUS to rent prediction, a task with long-term trends and seasonal fluctuations, using real-world datasets from Sydney and New York. Experimental results show that MARCUS achieves state-of-the-art performance, reducing MAE by 51.35% on Sydney and 12.62% on New York compared with the best baselines. Additional experiments, including an imputation-based ablation study and randomized additional-missingness analysis, further demonstrate the effectiveness of the proposed method.

cs.LG↗

arXiv2026

EVMbench: Evaluating AI Agents on Smart Contract Security

Smart contracts on public blockchains now manage large amounts of value, and vulnerabilities in these systems can lead to substantial losses. As AI agents become more capable at reading, writing, and running code, it is natural to ask how well they can already navigate this landscape, both in ways that improve security and in ways that might increase risk. We introduce EVMbench, an evaluation that measures the ability of agents to detect, patch, and exploit smart contract vulnerabilities. EVMbench draws on 117 curated vulnerabilities from 40 repositories and, in the most realistic setting, uses programmatic grading based on tests and blockchain state under a local Ethereum execution environment. We evaluate a range of frontier agents and find that they are capable of discovering and exploiting vulnerabilities end-to-end against live blockchain instances. We release code, tasks, and tooling to support continued measurement of these capabilities and future work on security.

cs.LG↗

arXiv2025

A low-cost, open-source maskless photolithography stepper for microfabrication

Photolithography is a key part of modern semiconductor process flows, and photolithography steppers have been used for decades to achieve precise patterning for device fabrication. However, these tools are often large and expensive, which restricts their use to industry and well-funded university laboratories. In this paper, we propose a $3000 maskless photolithography stepper that is affordable, open-source, and easy to assemble. The stepper, which uses a Digital Light Processing (DLP) projector as its optical engine, is able to achieve an optical resolution of under 2 microns. The stepper also features a motorized micropositioning system, which is able to align features with single-digit micron precision. A deep learning computer vision model is also used to achieve fine-grain alignment of patterns on a chip. These capabilities allow for the use of the stepper for microfabrication to produce micron-scale devices such as NMOS transistors.

physics.ins-det↗

arXiv2025

GRASP: Municipal Budget AI Chatbots for Enhancing Civic Engagement

There are a growing number of AI applications, but none tailored specifically to help residents answer their questions about municipal budget, a topic most are interested in but few have a solid comprehension of. In this research paper, we propose GRASP, a custom AI chatbot framework which stands for Generation with Retrieval and Action System for Prompts. GRASP provides more truthful and grounded responses to user budget queries than traditional information retrieval systems like general Large Language Models (LLMs) or web searches. These improvements come from the novel combination of a Retrieval-Augmented Generation (RAG) framework ("Generation with Retrieval") and an agentic workflow ("Action System"), as well as prompt engineering techniques, the incorporation of municipal budget domain knowledge, and collaboration with local town officials to ensure response truthfulness. During testing, we found that our GRASP chatbot provided precise and accurate responses for local municipal budget queries 78% of the time, while GPT-4o and Gemini were only accurate 60% and 35% of the time, respectively. GRASP chatbots greatly reduce the time and effort needed for the general public to get an intuitive and correct understanding of their town's budget, thus fostering greater communal discourse, improving government transparency, and allowing citizens to make more informed decisions.

cs.AI↗

arXiv2025

Building A Proof-Oriented Programmer That Is 64% Better Than GPT-4o Under Data Scarcity

Existing LMs struggle with proof-oriented programming due to data scarcity, which manifest in two key ways: (1) a lack of sufficient corpora for proof-oriented programming languages such as F*, and (2) the absence of large-scale, project-level proof-oriented implementations that can teach the model the intricate reasoning process when performing proof-oriented programming. We present the first on synthetic data augmentation for project level proof oriented programming for both generation and repair. Our method addresses data scarcity by synthesizing basic proof-oriented programming problems for proficiency in that language; incorporating diverse coding data for reasoning capability elicitation and creating new proofs and repair data within existing repositories. This approach enables language models to both synthesize and repair proofs for function- and repository-level code. We show that our fine-tuned 14B parameter model, PoPilot, can exceed the performance of the models that outperforms GPT-4o in project-level proof-oriented programming by 64% relative margin, and can improve GPT-4o's performance by 54% by repairing its outputs over GPT-4o's self-repair.

cs.CL↗

arXiv2024

AgentHarm: A Benchmark for Measuring Harmfulness of LLM Agents

The robustness of LLMs to jailbreak attacks, where users design prompts to circumvent safety measures and misuse model capabilities, has been studied primarily for LLMs acting as simple chatbots. Meanwhile, LLM agents -- which use external tools and can execute multi-stage tasks -- may pose a greater risk if misused, but their robustness remains underexplored. To facilitate research on LLM agent misuse, we propose a new benchmark called AgentHarm. The benchmark includes a diverse set of 110 explicitly malicious agent tasks (440 with augmentations), covering 11 harm categories including fraud, cybercrime, and harassment. In addition to measuring whether models refuse harmful agentic requests, scoring well on AgentHarm requires jailbroken agents to maintain their capabilities following an attack to complete a multi-step task. We evaluate a range of leading LLMs, and find (1) leading LLMs are surprisingly compliant with malicious agent requests without jailbreaking, (2) simple universal jailbreak templates can be adapted to effectively jailbreak agents, and (3) these jailbreaks enable coherent and malicious multi-step agent behavior and retain model capabilities. To enable simple and reliable evaluation of attacks and defenses for LLM-based agents, we publicly release AgentHarm at https://huggingface.co/datasets/ai-safety-institute/AgentHarm.

cs.LG↗

arXiv2024

$\textbf{Only-IF}$:Revealing the Decisive Effect of Instruction Diversity on Generalization

Understanding and accurately following instructions is critical for large language models (LLMs) to be effective across diverse tasks. In this work, we rigorously examine the key factors that enable models to generalize to unseen instructions, providing insights to guide the collection of data for instruction-tuning. Through controlled experiments, inspired by the Turing-complete Markov algorithm, we demonstrate that such generalization $\textbf{only emerges}$ when training data is diversified enough across semantic domains. Our findings also reveal that merely diversifying within limited domains fails to ensure robust generalization. In contrast, cross-domain data diversification, even under constrained data budgets, significantly enhances a model's adaptability. We further extend our analysis to real-world scenarios, including fine-tuning of $\textit{$\textbf{specialist}$}$ and $\textit{$\textbf{generalist}$}$ models. In both cases, we demonstrate that 1) better performance can be achieved by increasing the diversity of an established dataset while keeping the data size constant, and 2) when scaling up the data, diversifying the semantics of instructions is more effective than simply increasing the quantity of similar data. Our research provides important insights for dataset collation, particularly when optimizing model performance by expanding training data for both specialist and generalist scenarios. We show that careful consideration of data diversification is key: training specialist models with data extending beyond their core domain leads to significant performance improvements, while generalist models benefit from diverse data mixtures that enhance their overall instruction-following capabilities across a wide range of applications. Our results highlight the critical role of strategic diversification and offer clear guidelines for improving data quality.

cs.CL↗

arXiv2024

MAGICS: Adversarial RL with Minimax Actors Guided by Implicit Critic Stackelberg for Convergent Neural Synthesis of Robot Safety

While robust optimal control theory provides a rigorous framework to compute robot control policies that are provably safe, it struggles to scale to high-dimensional problems, leading to increased use of deep learning for tractable synthesis of robot safety. Unfortunately, existing neural safety synthesis methods often lack convergence guarantees and solution interpretability. In this paper, we present Minimax Actors Guided by Implicit Critic Stackelberg (MAGICS), a novel adversarial reinforcement learning (RL) algorithm that guarantees local convergence to a minimax equilibrium solution. We then build on this approach to provide local convergence guarantees for a general deep RL-based robot safety synthesis algorithm. Through both simulation studies on OpenAI Gym environments and hardware experiments with a 36-dimensional quadruped robot, we show that MAGICS can yield robust control policies outperforming the state-of-the-art neural safety synthesis methods.

cs.RO↗

arXiv2024

Tamper-Resistant Safeguards for Open-Weight LLMs

Rapid advances in the capabilities of large language models (LLMs) have raised widespread concerns regarding their potential for malicious use. Open-weight LLMs present unique challenges, as existing safeguards lack robustness to tampering attacks that modify model weights. For example, recent works have demonstrated that refusal and unlearning safeguards can be trivially removed with a few steps of fine-tuning. These vulnerabilities necessitate new approaches for enabling the safe release of open-weight LLMs. We develop a method, called TAR, for building tamper-resistant safeguards into open-weight LLMs such that adversaries cannot remove the safeguards even after hundreds of steps of fine-tuning. In extensive evaluations and red teaming analyses, we find that our method greatly improves tamper-resistance while preserving benign capabilities. Our results demonstrate that progress on tamper-resistance is possible, opening up a promising new avenue to improve the safety and security of open-weight LLMs.

cs.LG↗

arXiv2024

Exploring Parent-Child Perceptions on Safety in Generative AI: Concerns, Mitigation Strategies, and Design Implications

The widespread use of Generative Artificial Intelligence (GAI) among teenagers has led to significant misuse and safety concerns. To identify risks and understand parental controls challenges, we conducted a content analysis on Reddit and interviewed 20 participants (seven teenagers and 13 parents). Our study reveals a significant gap in parental awareness of the extensive ways children use GAI, such as interacting with character-based chatbots for emotional support or engaging in virtual relationships. Parents and children report differing perceptions of risks associated with GAI. Parents primarily express concerns about data collection, misinformation, and exposure to inappropriate content. In contrast, teenagers are more concerned about becoming addicted to virtual relationships with GAI, the potential misuse of GAI to spread harmful content in social groups, and the invasion of privacy due to unauthorized use of their personal data in GAI applications. The absence of parental control features on GAI platforms forces parents to rely on system-built controls, manually check histories, share accounts, and engage in active mediation. Despite these efforts, parents struggle to grasp the full spectrum of GAI-related risks and to perform effective real-time monitoring, mediation, and education. We provide design recommendations to improve parent-child communication and enhance the safety of GAI use.

cs.HC↗

arXiv2024

Improving Alignment and Robustness with Circuit Breakers

AI systems can take harmful actions and are highly vulnerable to adversarial attacks. We present an approach, inspired by recent advances in representation engineering, that interrupts the models as they respond with harmful outputs with "circuit breakers." Existing techniques aimed at improving alignment, such as refusal training, are often bypassed. Techniques such as adversarial training try to plug these holes by countering specific attacks. As an alternative to refusal training and adversarial training, circuit-breaking directly controls the representations that are responsible for harmful outputs in the first place. Our technique can be applied to both text-only and multimodal language models to prevent the generation of harmful outputs without sacrificing utility -- even in the presence of powerful unseen attacks. Notably, while adversarial robustness in standalone image recognition remains an open challenge, circuit breakers allow the larger multimodal system to reliably withstand image "hijacks" that aim to produce harmful content. Finally, we extend our approach to AI agents, demonstrating considerable reductions in the rate of harmful actions when they are under attack. Our approach represents a significant step forward in the development of reliable safeguards to harmful behavior and adversarial attacks.

cs.LG↗

arXiv2024

From Symbolic Tasks to Code Generation: Diversification Yields Better Task Performers

Instruction tuning -- tuning large language models on instruction-output pairs -- is a promising technique for making models better adapted to the real world. Yet, the key factors driving the model's capability to understand and follow instructions not seen during training remain under-explored. Our investigation begins with a series of synthetic experiments within the theoretical framework of a Turing-complete algorithm called Markov algorithm, which allows fine-grained control over the instruction-tuning data. Generalization and robustness with respect to the training distribution emerge once a diverse enough set of tasks is provided, even though very few examples are provided for each task. We extend these initial results to a real-world application scenario of code generation and find that a more diverse instruction set, extending beyond code-related tasks, improves the performance of code generation. Our observations suggest that a more diverse semantic space for instruction-tuning sets greatly improves the model's ability to follow instructions and perform tasks.

cs.CL↗

arXiv2024

Instruction Diversity Drives Generalization To Unseen Tasks

Instruction tuning -- fine-tuning a large language model (LLM) on pairs of instructions and desired outcomes -- is an approach that enables pre-trained language models to perform real-world tasks and follow human instructions. Its practical success depends on the model learning a broader set of instructions than those it was trained on. Yet the factors that determine model generalization to such \emph{unseen tasks} are not well understood. %To understand the driving factors of generalization, In this paper, we experiment with string rewrites, a symbolic task that serves as a building block for Turing complete Markov algorithms while allowing experimental control of "inputs" and "instructions". We investigate the trade-off between the number of instructions the model is trained on and the number of training samples provided for each instruction and observe that the diversity of the instruction set determines generalization. Generalization emerges once a diverse enough set of tasks is provided, even though very few examples are provided for each task. Instruction diversity also ensures robustness with respect to non-uniform distributions of instructions in the training set.

cs.CL↗

arXiv2021

Purely viscous acoustic propulsion of bimetallic rods

Synthetic microswimmers offer models for cell motility and their tunability makes them promising candidates for biomedical applications. Here we measure the acoustic propulsion of bimetallic micro-rods that, when trapped at the nodal plane of a MHz acoustic resonator, swim with speeds of up to 300 microns per second. While past acoustic streaming models predict speeds that are more than one order of magnitude smaller than our measurements, we demonstrate that the acoustic locomotion of the rods is driven by a viscous, non-reciprocal mechanism relying on shape anisotropy akin to that used by swimming cells and that reproduces our data with no adjustable parameters.

cond-mat.soft↗

arXiv2020

3D Pose Detection in Videos: Focusing on Occlusion

In this work, we build upon existing methods for occlusion-aware 3D pose detection in videos. We implement a two stage architecture that consists of the stacked hourglass network to produce 2D pose predictions, which are then inputted into a temporal convolutional network to produce 3D pose predictions. To facilitate prediction on poses with occluded joints, we introduce an intuitive generalization of the cylinder man model used to generate occlusion labels. We find that the occlusion-aware network is able to achieve a mean-per-joint-position error 5 mm less than our linear baseline model on the Human3.6M dataset. Compared to our temporal convolutional network baseline, we achieve a comparable mean-per-joint-position error of 0.1 mm less at reduced computational cost.

cs.CV↗

arXiv2020

A 6G White Paper on Connectivity for Remote Areas

In many places all over the world rural and remote areas lack proper connectivity that has led to increasing digital divide. These areas might have low population density, low incomes, etc., making them less attractive places to invest and operate connectivity networks. 6G could be the first mobile radio generation truly aiming to close the digital divide. However, in order to do so, special requirements and challenges have to be considered since the beginning of the design process. The aim of this white paper is to discuss requirements and challenges and point out related, identified research topics that have to be solved in 6G. This white paper first provides a generic discussion, shows some facts and discusses targets set in international bodies related to rural and remote connectivity and digital divide. Then the paper digs into technical details, i.e., into a solutions space. Each technical section ends with a discussion and then highlights identified 6G challenges and research ideas as a list.

eess.SP↗

arXiv2015

Discrete All-Pay Bidding Games

In an all-pay auction, only one bidder wins but all bidders must pay the auctioneer. All-pay bidding games arise from attaching a similar bidding structure to traditional combinatorial games to determine which player moves next. In contrast to the established theory of single-pay bidding games, optimal play involves choosing bids from some probability distribution that will guarantee a minimum probability of winning. In this manner, all-pay bidding games wed the underlying concepts of economic and combinatorial games. We present several results on the structures of optimal strategies in these games. We then give a fast algorithm for computing such strategies for a large class of all-pay bidding games. The methods presented provide a framework for further development of the theory of all-pay bidding games.

cs.GT↗