This is a experimental project. Feel free to send feedback!

Thesis Tide

Thesis Tide ranks papers based on their relevance to the fields, with the goal of making it easier to find the most relevant papers. It uses AI to analyze the content of papers and rank them!

The Redshift-Space Momentum Power Spectrum III: measuring the growth rate from the SDSSv survey using auto- and cross- power spectrum of the galaxy density and momentum fields

The large-scale structure of the Universe and its evolution over time contains an abundance of cosmological information. One way to unlock this is by measuring the density and momentum power spectrum ...

Useful Fields:

The article presents a novel methodology for measuring the growth rate of structures in the universe by integrating momentum fields with density fields, utilizing advanced statistical techniques. The use of SDSS data enhances its applicability and robustness. By improving parameter estimation through cross-power spectrum analysis, the findings could significantly influence future cosmological studies related to the evolution of large-scale structures.

OpenGeMM: A High-Utilization GeMM Accelerator Generator with Lightweight RISC-V Control and Tight Memory Coupling

Deep neural networks (DNNs) face significant challenges when deployed on resource-constrained extreme edge devices due to their computational and data-intensive nature. While standalone accelerators t...

Useful Fields:

The article presents an innovative solution for enhancing computational efficiency in DNNs on edge devices, addressing key challenges of flexibility, programmability, and high utilization. The use of a RISC-V architecture combined with a configurable accelerator showcases methodological rigor and relevance to current technological needs in edge computing. Its open-source nature promotes further research and adaptation in the field, enhancing its impact potential.

Riesz-Schur transforms

We investigate nontrigonometric forms of Riesz transforms in the context of Schur multipliers. This refines Grothendieck-Haagerup's endpoint criterion with a new condition for the Schatten p-bound...

Useful Fields:

The article presents significant advancements in the field of harmonic analysis by exploring nontrigonometric forms of Riesz transforms, which is a novel approach. The refinement of established criteria, along with the simpler methodology, shows promise for broader applicability and further exploration. The theoretical contributions made in connection with Schur multipliers and the recovery of dimension-free estimates indicate a high potential for future research development.

Rethinking Weight-Averaged Model-merging

Weight-averaged model-merging has emerged as a powerful approach in deep learning, capable of enhancing model performance without fine-tuning or retraining. However, the underlying mechanisms that exp...

Useful Fields:

The article presents a novel exploration of weight-averaged model-merging, addressing a critical gap in understanding its mechanisms. Its methodological rigor in examining multiple perspectives contributes significantly to the field of deep learning. The insights offered may lead to improved practices in model development and ensemble methods, enhancing the applicability and effectiveness of machine learning systems.

Generalized Cauchy-Riemann equations and relevant PDE

Here we give a survey of consequences from the theory of the Beltrami equations in the complex plane $\mathbb C$ to generalized Cauchy-Riemann equations $\nabla v = B \nabla u$ in the ...

Useful Fields:

The article presents a comprehensive survey of significant mathematical concepts integrating the Beltrami equation and generalized Cauchy-Riemann equations, which are foundational in potential theory and fluid mechanics. Its focus on key boundary value problems and existence theorems ensures its relevance in both theoretical and applied contexts. The potential applications in hydro-mechanics offer interdisciplinary significance, contributing to engineering and physical sciences.

Approximation rates of entropic maps in semidiscrete optimal transport

Entropic optimal transport offers a computationally tractable approximation to the classical problem. In this note, we study the approximation rate of the entropic optimal transport map (in approachin...

Useful Fields:

The article presents significant advancements in the theory of entropic optimal transport, particularly by improving the approximation rates with respect to established norms. The results are relevant not only for theoretical implications but also for practical applications in computational optimal transport, which can affect various domains, including statistics and machine learning. The rigorous mathematical analysis and novel findings enhance the literature in the field and have the potential to inspire further research on optimal transport methods and approximations.

Crunch from AdS bubble collapse in unbounded potentials

We consider a scalar field theory with a Minkowski false vacuum and an unbounded (or very deep) true vacuum. We show compelling evidence that an AdS bubble of vanishing total energy, embedded in asymp...

Useful Fields:

This article presents novel insights into the behavior of scalar fields in asymptotically flat spacetimes, particularly regarding bubble collapse and singularity formation. Its implications for understanding curvature singularities and trapped surfaces can significantly advance theoretical physics, especially in the context of cosmology and quantum gravity.

Simulating the two-dimensional $t-J$ model at finite doping with neural quantum states

Simulating large, strongly interacting fermionic systems remains a major challenge for existing numerical methods. In this work, we present, for the first time, the application of neural quantum state...

Useful Fields:

This article presents a novel approach by employing neural quantum states for simulating the $t-J$ model, a significant advancement in the field of condensed matter physics. It not only addresses a critical challenge in computational physics but also showcases strong methodological rigor through comparisons with established methods like MPS. The ability to probe low-energy physics over a range of doping levels adds substantial applicability, especially for understanding emergent phenomena in strongly correlated systems.

VeriGraph: Scene Graphs for Execution Verifiable Robot Planning

Recent advancements in vision-language models (VLMs) offer potential for robot task planning, but challenges remain due to VLMs' tendency to generate incorrect action sequences. To address these l...

Useful Fields:

The article presents a fruitful integration of vision-language models and verifies actions using scene graphs, addressing a critical issue in robotic planning. Its methodological rigour is demonstrated by substantial performance improvements in task completion rates. The work is novel, as it bridges a gap in current robot planning approaches by providing a mechanism for verifying action feasibility through a structured representation of the environment, which could inspire further research into hybrid VLM applications.

BiDense: Binarization for Dense Prediction

Dense prediction is a critical task in computer vision. However, previous methods often require extensive computational resources, which hinders their real-world application. In this paper, we propose...

Useful Fields:

The article presents a novel approach to dense prediction in computer vision, addressing crucial limitations in computational resource requirements of previous methods. The innovative techniques, DAB and CFB, enhance the efficiency and accuracy of binary neural networks (BNNs). The scalability of these methods suggests potential for widespread application and relevance in both academic and practical contexts, making it a significant contribution to the field.

Shadow and accretion disk images of Kerr-Newman black hole in modified gravity theory

In this paper, we study the shadow and images of the accretion disk of Kerr-Newman (KN) black hole (BH) in modified gravity (MOG) theory by using backward ray-tracing method. And, the influence of spi...

Useful Fields:

This article tackles the intriguing topic of Kerr-Newman black holes within the modified gravity context, contributing original insights about the effects of the MOG parameter on black hole imagery. Its use of backward ray-tracing adds methodological robustness, and the focus on observables makes the findings applicable for both theoretical and observational astrophysics. The clear implications for gravitational lensing and astrophysical observations enhance its potential impact and relevance to future studies in black hole physics.

MOT FCG++: Enhanced Representation of Spatio-temporal Motion and Appearance Features

The goal of multi-object tracking (MOT) is to detect and track all objects in a scene across frames, while maintaining a unique identity for each object. Most existing methods rely on the spatial-temp...

Useful Fields:

The article presents a novel approach to enhance motion and appearance feature representation in multi-object tracking (MOT), addressing critical performance issues in existing methods. Its methodological rigor, demonstrated by the improvements achieved on multiple benchmarks (MOT17, MOT20, and DanceTrack), positions it as a significant contribution to the field. The novelty of introducing Diagonal Modulated GIoU and dynamic appearance representation additionally signals a potential shift in future research directions within MOT.

Pro-Prophet: A Systematic Load Balancing Method for Efficient Parallel Training of Large-scale MoE Models

The size of deep learning models has been increasing to enhance model quality. The linear increase in training computation budget with model size means that training an extremely large-scale model is ...

Useful Fields:

The article presents a novel systematic load balancing method, Pro-Prophet, which specifically addresses critical inefficiencies in the training of large-scale Mixture of Expert models. Its potential to significantly improve training throughput and address load imbalance positions it as a valuable contribution to the field. The methodological rigor, evidenced by experimental validation showing substantial performance improvements, further underscores its relevance. Moreover, the focus on system-level solutions ensures its applicability in real-world scenarios, making the research highly influential for subsequent studies and applications in large-scale model training.

Instruction-Guided Editing Controls for Images and Multimedia: A Survey in LLM era

The rapid advancement of large language models (LLMs) and multimodal learning has transformed digital content creation and manipulation. Traditional visual editing tools require significant expertise,...

Useful Fields:

This article provides a comprehensive survey on a cutting-edge topic at the intersection of large language models and multimodal editing, which is highly relevant given the current trends in AI and digital content creation. Its breadth of coverage (over 100 publications) and focus on accessibility and practical applications underscore its methodological rigor and broad applicability. Furthermore, it identifies key challenges and encourages future research, enhancing its potential impact.

High-Speed Cornering Control and Real-Vehicle Deployment for Autonomous Electric Vehicles

Executing drift maneuvers during high-speed cornering presents significant challenges for autonomous vehicles, yet offers the potential to minimize turning time and enhance driving dynamics. While rei...

Useful Fields:

This article introduces a novel approach to real-world implementation of high-speed cornering control in autonomous electric vehicles, systematically combining advanced reinforcement learning techniques with practical vehicle control frameworks. The integration of Bezier-based trajectory optimization, TD3, and MPC demonstrates methodological rigor and applicability, while providing a compelling case for their real-world deployment. The validation through real-vehicle tests adds substantial credibility to the claims made, enhancing the relevance of this research for advancing autonomous vehicle technology and control strategies.

Interval Multiplicities of Persistence Modules

For any persistence module $M$ over a finite poset $\mathbf{P}$ , and any interval $I$ in $\mathbf{P}$ , we give a formula of the multiplicity $d_M(V_I)$ of the i...

Useful Fields:

The article presents a significant advancement in the understanding of persistence modules through the introduction of a new formula for multiplicities, which is both novel and applicable to ongoing research in topological data analysis. The proposed method for determining interval-decomposability expands the toolkit available for mathematicians and computer scientists in the field. Theoretical insights coupled with operational methodologies enhance its potential impact.

On the Density of naturals $n$ coprime to $\lfloor P(n) \rfloor$ for certain Classes of Polynomials

We obtain asymptotic bounds on the number of natural numbers less than $X$ satisfying $\gcd{\left(n,\lfloor P(n) \rfloor\right)}=1$ , under some diophantine conditions on the coefficien...

Useful Fields:

The article presents a novel result concerning the density of natural numbers coprime to a floor function of a polynomial, which bridges the gap between number theory and analysis. The asymptotic bounds provided under specific conditions add methodological rigor, and the identification of the density as a specific value (1/ζ(2)) opens pathways for further research. However, the applicability may be somewhat limited to a niche area within analytic number theory.

Hamiltonian stationary Lagrangian surfaces with harmonic mean curvature in complex space forms

We completely classify Hamiltonian stationary Lagrangian surfaces with harmonic mean curvature and constant curvature in complex space forms.

Useful Fields:

The article addresses a complex mathematical problem by classifying Hamiltonian stationary Lagrangian surfaces, which is significant as it contributes to the understanding of geometry and offers potential applications in physics (e.g., in string theory). The work appears to employ rigorous theoretical methods, promising robustness. The novel approach to understanding curvature properties in complex space forms is particularly valuable for future research in differential geometry.

Controlling Diversity at Inference: Guiding Diffusion Recommender Models with Targeted Category Preferences

Diversity control is an important task to alleviate bias amplification and filter bubble problems. The desired degree of diversity may fluctuate based on users' daily moods or business strategies....

Useful Fields:

The article presents a novel approach (D3Rec) to a significant issue in recommender systems—controlling diversity during inference. Its methodological innovation allows for dynamic adjustments to recommendations based on user preferences, which is a substantial improvement over existing static methods. The extensive experiments further support its effectiveness, highlighting both robustness and applicability across various scenarios, thus presenting high potential for future research in recommender systems and related fields.

Optical Tweezers with AC Dielectric Levitation: A Powerful Approach to Microparticle Manipulation

Optical tweezers, with their high precision, dynamic control, and non-invasiveness, are increasingly important in scientific research and applications at the micro and nano scales. However, manipulati...

Useful Fields:

The article presents a novel and multipurpose technique that integrates AC dielectric levitation with optical tweezers, effectively overcoming existing limitations in microparticle manipulation. The robust experimental validation alongside finite element simulations enhances methodological rigor, suggesting that this method could significantly advance research in the field of micromanipulation.