arXiv subjects

Michael Zhang

Publications and source records attributed to Michael Zhang.

At least 37 records · Page 2Linked to original sources

arXiv2025

Self-Interest and Systemic Benefits: Emergence of Collective Rationality in Mixed Autonomy Traffic Through Deep Reinforcement Learning

Autonomous vehicles (AVs) are expected to be commercially available in the near future, leading to mixed autonomy traffic consisting of both AVs and human-driven vehicles (HVs). Although numerous studies have shown that AVs can be deployed to benefit the overall traffic system performance by incorporating system-level goals into their decision making, it is not clear whether the benefits still exist when agents act out of self-interest -- a trait common to all driving agents, both human and autonomous. This study aims to understand whether self-interested AVs can bring benefits to all driving agents in mixed autonomy traffic systems. The research is centered on the concept of collective rationality (CR). This concept, originating from game theory and behavioral economics, means that driving agents may cooperate collectively even when pursuing individual interests. Our recent research has proven the existence of CR in an analytical game-theoretical model and empirically in mixed human-driven traffic. In this paper, we demonstrate that CR can be attained among driving agents trained using deep reinforcement learning (DRL) with a simple reward design. We examine the extent to which self-interested traffic agents can achieve CR without directly incorporating system-level objectives. Results show that CR consistently emerges in various scenarios, which indicates the robustness of this property. We also postulate a mechanism to explain the emergence of CR in the microscopic and dynamic environment and verify it based on simulation evidence. This research suggests the possibility of leveraging advanced learning methods (such as federated learning) to achieve collective cooperation among self-interested driving agents in mixed-autonomy systems.

cs.LG↗

arXiv2025

MobilityGPT: Enhanced Human Mobility Modeling with a GPT model

Generative models have shown promising results in capturing human mobility characteristics and generating synthetic trajectories. However, it remains challenging to ensure that the generated geospatial mobility data is semantically realistic, including consistent location sequences, and reflects real-world characteristics, such as constraining on geospatial limits. We reformat human mobility modeling as an autoregressive generation task to address these issues, leveraging the Generative Pre-trained Transformer (GPT) architecture. To ensure its controllable generation to alleviate the above challenges, we propose a geospatially-aware generative model, MobilityGPT. We propose a gravity-based sampling method to train a transformer for semantic sequence similarity. Then, we constrained the training process via a road connectivity matrix that provides the connectivity of sequences in trajectory generation, thereby keeping generated trajectories in geospatial limits. Lastly, we proposed to construct a preference dataset for fine-tuning MobilityGPT via Reinforcement Learning from Trajectory Feedback (RLTF) mechanism, which minimizes the travel distance between training and the synthetically generated trajectories. Experiments on real-world datasets demonstrate MobilityGPT's superior performance over state-of-the-art methods in generating high-quality mobility trajectories that are closest to real data in terms of origin-destination similarity, trip length, travel radius, link, and gravity distributions. We release the source code and reference links to datasets at https://github.com/ammarhydr/MobilityGPT.

cs.LG↗

arXiv2025

Design, Implementation and Evaluation of a Novel Programming Language Topic Classification Workflow

As software systems grow in scale and complexity, understanding the distribution of programming language topics within source code becomes increasingly important for guiding technical decisions, improving onboarding, and informing tooling and education. This paper presents the design, implementation, and evaluation of a novel programming language topic classification workflow. Our approach combines a multi-label Support Vector Machine (SVM) with a sliding window and voting strategy to enable fine-grained localization of core language concepts such as operator overloading, virtual functions, inheritance, and templates. Trained on the IBM Project CodeNet dataset, our model achieves an average F1 score of 0.90 across topics and 0.75 in code-topic highlight. Our findings contribute empirical insights and a reusable pipeline for researchers and practitioners interested in code analysis and data-driven software engineering.

cs.SE↗

arXiv2025

Reverse Engineering User Stories from Code using Large Language Models

User stories are essential in agile development, yet often missing or outdated in legacy and poorly documented systems. We investigate whether large language models (LLMs) can automatically recover user stories directly from source code and how prompt design impacts output quality. Using 1,750 annotated C++ snippets of varying complexity, we evaluate five state-of-the-art LLMs across six prompting strategies. Results show that all models achieve, on average, an F1 score of 0.8 for code up to 200 NLOC. Our findings show that a single illustrative example enables the smallest model (8B) to match the performance of a much larger 70B model. In contrast, structured reasoning via Chain-of-Thought offers only marginal gains, primarily for larger models.

cs.SE↗

arXiv2025

MMORE: Massive Multimodal Open RAG & Extraction

We introduce MMORE, an open-source pipeline for Massive Multimodal Open RetrievalAugmented Generation and Extraction, designed to ingest, transform, and retrieve knowledge from heterogeneous document formats at scale. MMORE supports more than fifteen file types, including text, tables, images, emails, audio, and video, and processes them into a unified format to enable downstream applications for LLMs. The architecture offers modular, distributed processing, enabling scalable parallelization across CPUs and GPUs. On processing benchmarks, MMORE demonstrates a 3.8-fold speedup over single-node baselines and 40% higher accuracy than Docling on scanned PDFs. The pipeline integrates hybrid dense-sparse retrieval and supports both interactive APIs and batch RAG endpoints. Evaluated on PubMedQA, MMORE-augmented medical LLMs improve biomedical QA accuracy with increasing retrieval depth. MMORE provides a robust, extensible foundation for deploying task-agnostic RAG systems on diverse, real-world multimodal data. The codebase is available at https://github.com/swiss-ai/mmore.

cs.SE↗

arXiv2025

Stringent Upper Bounds on Atmospheric Mass Loss from Three Neptune-Sized Planets in the TOI-4010 System

Photoevaporative models predict that the lower edge of the Neptune desert is sculpted by atmospheric mass loss. However, the stellar high energy fluxes that power hydrodynamic escape and set predicted mass loss rates can be uncertain by multiple orders of magnitude. These uncertainties can be bypassed by studying mass loss for planets within the same system, as they have effectively undergone scaled versions of the same irradiation history. The TOI-4010 system is an ideal test case for mass loss models, as it contains three Neptune-sized planets with planet b located in the `Neptune desert', planet c in the `Neptune ridge', and planet d in the `Neptune savanna'. Using Keck/NIRSPEC, we measured the metastable helium transit depths of all three planets in order to search for evidence of atmospheric escape. We place upper bounds on the excess helium absorption of 1.23\%, 0.81\%, and 0.87\% at 95\% confidence for TOI-4010~b, c and d respectively. We fit our transmission spectra with Parker wind models and find that this corresponds to 95th-percentile upper limits of $10^{10.17}$g~s$^{-1}$, $10^{10.53}$g~s$^{-1}$, and $10^{10.50}$g~s$^{-1}$ on the mass loss rates of TOI-4010~b, c, and d respectively. Our non-detections are inconsistent with expectations from one-dimensional hydrodynamic models for solar composition atmospheres. We consider potential reductions in signal from a decreased host star XUV luminosity, planetary magnetic fields, enhanced atmospheric metallicities, and fractionation, and explore the implications of our measurements for the past evaporation histories of all three planets.

astro-ph.EP↗

arXiv2025

The JWST Rocky Worlds DDT Program reveals GJ 3929b to likely be a bare rock

We report first results from the JWST Rocky Worlds Director's Discretionary Time program. Two secondary eclipses of the terrestrial exoplanet GJ 3929b were recently observed using MIRI photometric imaging at 15 um. We present a reduction of these data using the updated SPARTA pipeline. We also refine the planet mass, radius, and predicted time of secondary eclipse using a new sector of TESS data and new, high-precision radial velocities from the MAROON-X spectrograph. For the two JWST observations, we recover secondary eclipse depths of 177+47-45ppm and 143+34-35ppm at times consistent with a nearly circular orbit, as expected from the radial velocity data. A joint fit of the two visits yields a dayside brightness temperature Tp,dayside = 782+/-79K for GJ 3929b, which is consistent with the maximum brightness temperature Tmax = 737+/-14K for a bare, black rock (i.e., assuming zero Bond albedo and no heat redistribution). These results rule out CO2-rich atmospheres thicker than 100mbar at >3sigma, suggesting that GJ 3929b has lost any significant secondary atmosphere. The radial velocity data also indicate two additional non-transiting planets in the system: a previously-identified planet in a 15.0d orbit, and a newly-identified planet candidate in a 6.1d orbit.

astro-ph.EP↗

arXiv2025

Population-level Hypothesis Testing with Rocky Planet Emission Data: A Tentative Trend in the Brightness Temperatures of M-Earths

Determining which rocky exoplanets have atmospheres, and why, is a key goal for the James Webb Space Telescope. So far, emission observations of individual rocky exoplanets orbiting M stars (M-Earths) have not provided definitive evidence for atmospheres. Here, we synthesize emission data for M-Earths and find a trend in measured brightness temperatures (ratioed to its theoretical maximum value) as a function of instellation. However, the statistical evidence of this trend is dependent on the choice of stellar model, and we consider its identification tentative. We show that this trend can be explained by either the onset of thin/tenuous (<1 bar) CO2-rich atmospheres on colder worlds, or a population of bare rocks with stronger space weathering and/or coarser regolith on closer-in worlds. Such grain coarsening may be caused by sintering near the melting point of rock or frequent volcanic resurfacing. Furthermore, we highlight considerations when testing rocky planet hypotheses at the population level, including the choice of instrument, stellar modeling, and how brightness temperatures are derived. We also find that fresh (unweathered) fine-grained surfaces can serve as a false positive to the detection of moderate atmospheric heat redistribution through eclipse observations. However, we argue that such surfaces are unlikely given the ubiquity of space weathering in the Solar System, the low albedo of Solar System airless bodies, and the high stellar wind environments of M-Earths. Emission data from a larger sample of M-Earths will be able to confirm or reject this tentative trend and diagnose its cause through spectral characterization.

astro-ph.EP↗

arXiv2025

Enhanced Vehicle Speed Detection Considering Lane Recognition Using Drone Videos in California

The increase in vehicle numbers in California, driven by inadequate transportation systems and sparse speed cameras, necessitates effective vehicle speed detection. Detecting vehicle speeds per lane is critical for monitoring High-Occupancy Vehicle (HOV) lane speeds, distinguishing between cars and heavy vehicles with differing speed limits, and enforcing lane restrictions for heavy vehicles. While prior works utilized YOLO (You Only Look Once) for vehicle speed detection, they often lacked accuracy, failed to identify vehicle lanes, and offered limited or less practical classification categories. This study introduces a fine-tuned YOLOv11 model, trained on almost 800 bird's-eye view images, to enhance vehicle speed detection accuracy which is much higher compare to the previous works. The proposed system identifies the lane for each vehicle and classifies vehicles into two categories: cars and heavy vehicles. Designed to meet the specific requirements of traffic monitoring and regulation, the model also evaluates the effects of factors such as drone height, distance of Region of Interest (ROI), and vehicle speed on detection accuracy and speed measurement. Drone footage collected from Northern California was used to assess the proposed system. The fine-tuned YOLOv11 achieved its best performance with a mean absolute error (MAE) of 0.97 mph and mean squared error (MSE) of 0.94 $\text{mph}^2$, demonstrating its efficacy in addressing challenges in vehicle speed detection and classification.

cs.CV↗

arXiv2025

Fast Estimation of Globally Optimal Independent Contact Regions for Robust Grasping and Manipulation

This work presents a fast anytime algorithm for computing globally optimal independent contact regions (ICRs). ICRs are regions such that one contact within each region enables a valid grasp. Locations of ICRs can provide guidance for grasp and manipulation planning, learning, and policy transfer. However, ICRs for modern applications have been little explored, in part due to the expense of computing them, as they have a search space exponential in the number of contacts. We present a divide and conquer algorithm based on incremental n-dimensional Delaunay triangulation that produces results with bounded suboptimality in times sufficient for real-time planning. This paper presents the base algorithm for grasps where contacts lie within a plane. Our experiments show substantial benefits over competing grasp quality metrics and speedups of 100X and more for competing approaches to computing ICRs. We explore robustness of a policy guided by ICRs and outline a path to general 3D implementation. Code will be released on publication to facilitate further development and applications.

cs.RO↗

arXiv2025

Characterising the Atmosphere of 55 Cancri e: 1D Forward Model Grid for Current and Future JWST Observations

Recent JWST observations with NIRCam and MIRI of the ultra-short-period super-Earth 55 Cancri e indicate a possible volatile atmosphere surrounding the planet. Previous analysis of the NIRCam spectra suggested potential absorption features from CO2 or CO and significant sub-weekly variability. The MIRI low-resolution spectrum does not contain substantial features but was found to be consistent with effective heat redistribution models. In this work, we computed a grid of over 25000 self-consistent 1D forward models incorporating H-N-O-C-S-P-Si-Ti equilibrium chemistry and assessed plausible atmospheric compositions based on the current JWST data. Despite exhaustive analysis, the composition and properties of the atmosphere remain elusive. While our results statistically favour a global, hydrogen-free, nitrogen-dominated atmosphere enriched in PO and CO2, various alternative compositions, including H2O-,CO-, PH3-, or Si-bearing remain viable explanations. Unconstrained heat redistribution efficiency and absolute NIRCam flux are among the largest sources of uncertainty in our analysis. We also find that the heat redistribution factor and surface pressure are highly degenerate with atmospheric composition, and that these parameters cannot be independently constrained using current JWST observations. Furthermore, we show that the observed variability may arise from dynamic interactions between the atmosphere and an underlying magma ocean, driving rapid shifts in atmospheric chemistry and thermal emission. Our results highlight the importance of using self-consistent forward models when analysing novel JWST spectra with limited signal-to-noise ratios -- such as those of 55 Cancri e -- as it allows for a more comprehensive evaluation of potential atmospheric scenarios while also being less sensitive to subtle spectral differences than retrievals...

astro-ph.EP↗

arXiv2025

Guiding Data Collection via Factored Scaling Curves

Generalist imitation learning policies trained on large datasets show great promise for solving diverse manipulation tasks. However, to ensure generalization to different conditions, policies need to be trained with data collected across a large set of environmental factor variations (e.g., camera pose, table height, distractors) $-$ a prohibitively expensive undertaking, if done exhaustively. We introduce a principled method for deciding what data to collect and how much to collect for each factor by constructing factored scaling curves (FSC), which quantify how policy performance varies as data scales along individual or paired factors. These curves enable targeted data acquisition for the most influential factor combinations within a given budget. We evaluate the proposed method through extensive simulated and real-world experiments, across both training-from-scratch and fine-tuning settings, and show that it boosts success rates in real-world tasks in new environments by up to 26% over existing data-collection strategies. We further demonstrate how factored scaling curves can effectively guide data collection using an offline metric, without requiring real-world evaluation at scale.

cs.RO↗

arXiv2025

The radiative effects of photochemical hazes on the atmospheric circulation and phase curves of sub-Neptunes

Measuring the atmospheric composition of hazy sub-Neptunes like GJ~1214b through transmission spectroscopy is difficult because of the degeneracy between mean molecular weight and haziness. It has been proposed that phase curve observations can break this degeneracy because of the relationship between mean molecular weight (MMW) and phase curve amplitude. However, photochemical hazes can strongly affect phase curve amplitudes as well. We present a large set of GCM simulations of the sub-Neptune GJ~1214b that include photochemical hazes with varying atmospheric composition, haze opacity and haze optical properties. In our simulations, photochemical hazes cause temperature changes of up to 200~K, producing thermal inversions and cooling deeper regions. This results in increased phase curve amplitudes and adds a considerable scatter to the phase curve amplitude--metallicity relationship. However, we find that if the haze production rate is high enough to significantly alter the phase curve, the secondary eclipse spectrum will exhibit either emission features or strongly muted absorption features. Thus, the combination of a white-light phase curve and a secondary eclipse spectrum can successfully distinguish between a hazy, lower MMW and a clear, high MMW scenario.

astro-ph.EP↗

arXiv2025

Clouds and Hazes in GJ 1214b's Metal-Rich Atmosphere

The sub-Neptune GJ 1214b has an infamously flat transmission spectrum, likely due to thick aerosols in its atmosphere. A recent JWST MIRI spectroscopic phase curve of GJ 1214 b added to this picture, suggesting a highly reflective and metal-rich atmosphere. Using a 3D General Circulation Model with both photochemical hazes and condensate clouds, we characterize how different aerosol types affect the atmospheric structure of GJ 1214 b and manifest in its spectroscopic phase curve. Additionally, we reanalyze the original GJ 1214 b JWST phase curve. The reanalysis shows a hotter nightside, similar dayside temperature, and a lower, but still elevated, Bond albedo (0.42 +/- 0.11) than the original results. We find that a scenario with both clouds and hazes is most consistent with the JWST phase curve. Reflective clouds or hazes are needed to explain the large Bond albedo, and hazes or a super-solar metallicity help account for the several hundred Kelvin day-night temperature difference measured by the phase curve.

astro-ph.EP↗

arXiv2025

Simple linear attention language models balance the recall-throughput tradeoff

Recent work has shown that attention-based language models excel at recall, the ability to ground generations in tokens previously seen in context. However, the efficiency of attention-based models is bottle-necked during inference by the KV-cache's aggressive memory consumption. In this work, we explore whether we can improve language model efficiency (e.g. by reducing memory consumption) without compromising on recall. By applying experiments and theory to a broad set of architectures, we identify a key tradeoff between a model's state size and recall ability. We show that efficient alternatives to attention (e.g. H3, Mamba, RWKV) maintain a fixed-size recurrent state, but struggle at recall. We propose BASED a simple architecture combining linear and sliding window attention. By varying BASED window size and linear attention feature dimension, we can dial the state size and traverse the pareto frontier of the recall-memory tradeoff curve, recovering the full quality of attention on one end and the small state size of attention-alternatives on the other. We train language models up to 1.3b parameters and show that BASED matches the strongest sub-quadratic models (e.g. Mamba) in perplexity and outperforms them on real-world recall-intensive tasks by 6.22 accuracy points. Implementations of linear attention are often less efficient than optimized standard attention implementations. To make BASED competitive, we develop IO-aware algorithms that enable 24x higher throughput on language generation than FlashAttention-2, when generating 1024 tokens using 1.3b parameter models. Code for this work is provided at: https://github.com/HazyResearch/based.

cs.CL↗

arXiv2025

Constraining atmospheric composition from the outflow: helium observations reveal the fundamental properties of two planets straddling the radius gap

TOI-836 is a ~2-3 Gyr K dwarf with an inner super Earth ($R=1.7 R_\oplus$, $P=3.8$ d) and an outer mini Neptune ($R=2.6 R_\oplus$, $P=8.6$ d). JWST/NIRSpec 2.8--5.2 $μ$m transmission spectra are flat for both planets. We present Keck/NIRSPEC observations of escaping helium for super-Earth b, which shows no excess absorption in the 1083 nm triplet to deep limits (<0.2%), and mini-Neptune c, which shows strong (0.7%) excess absorption in both visits. These results demonstrate that planet c retains at least some primordial atmosphere, while planet b is consistent with having lost its entire primordial envelope. Self-consistent 1D radiative-hydrodynamic models of planet c reveal that the helium excess absorption signal is highly sensitive to metallicity: its equivalent width collapses by a factor of 13 as metallicity increases from 10x to 100x solar, and by a further factor of 12 as it increases to 200x solar. The observed equivalent width is 88% the model prediction for 100x metallicity, suggesting an atmospheric metallicity similar to K2-18b and TOI-270d, the first two mini-Neptunes with detected absorption features in JWST transmission spectra. We highlight the helium triplet as a potentially powerful probe of atmospheric composition, with complementary strengths and weaknesses to atmospheric retrievals. The main strength is its extreme sensitivity to metallicity in the scientifically significant range of 10--200x solar, and the main weakness is the enormous model uncertainties in outflow suppression and confinement mechanisms, such as magnetic fields and stellar winds, which can suppress the signal by at least a factor of ~several.

astro-ph.EP↗

arXiv2025

LoLCATs: On Low-Rank Linearizing of Large Language Models

Recent works show we can linearize large language models (LLMs) -- swapping the quadratic attentions of popular Transformer-based LLMs with subquadratic analogs, such as linear attention -- avoiding the expensive pretraining costs. However, linearizing LLMs often significantly degrades model quality, still requires training over billions of tokens, and remains limited to smaller 1.3B to 7B LLMs. We thus propose Low-rank Linear Conversion via Attention Transfer (LoLCATs), a simple two-step method that improves LLM linearizing quality with orders of magnitudes less memory and compute. We base these steps on two findings. First, we can replace an LLM's softmax attentions with closely-approximating linear attentions, simply by training the linear attentions to match their softmax counterparts with an output MSE loss ("attention transfer"). Then, this enables adjusting for approximation errors and recovering LLM quality simply with low-rank adaptation (LoRA). LoLCATs significantly improves linearizing quality, training efficiency, and scalability. We significantly reduce the linearizing quality gap and produce state-of-the-art subquadratic LLMs from Llama 3 8B and Mistral 7B v0.1, leading to 20+ points of improvement on 5-shot MMLU. Furthermore, LoLCATs does so with only 0.2% of past methods' model parameters and 0.4% of their training tokens. Finally, we apply LoLCATs to create the first linearized 70B and 405B LLMs (50x larger than prior work). When compared with prior approaches under the same compute budgets, LoLCATs significantly improves linearizing quality, closing the gap between linearized and original Llama 3.1 70B and 405B LLMs by 77.8% and 78.1% on 5-shot MMLU.

cs.LG↗

arXiv2025

Analysis of a platooned car-following model with different inter-platoon communication levels

Despite growing interest in vehicle platooning research, the effect of communication capability between platoons is not investigated to a depth of depth. In this paper, we extend a single-platoon car-following (CF) model to multi-platoon CF models for connected and autonomous vehicles (CAVs) with different inter-platoon communication capabilities. Specifically, we consider forward and backward connection availabilities with delays between platoons. Using linear stability analysis, we discovered that for identical platoons, stability increases with platoon size and connection availabilities and decreases exponentially with large delay. With maximum acceleration and emergency braking constraints integrated into the models, we performed simulations for various cases of CAV platoons and mixed autonomy with human-driven vehicles (HDVs). The simulation results for CAV platoons are consistent with theoretical analysis. The mixed autonomy experiments demonstrate that in the ring road scenario, CAV platoons can stabilize HDVs without adaptions, and the effect of distribution is marginal. Overall, this paper provides valuable insights for designing vehicle-to-vehicle (V2V) communications and managing mixed traffic scenarios.

eess.SY↗