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!

AI in Support of Diversity and Inclusion

In this paper, we elaborate on how AI can support diversity and inclusion and exemplify research projects conducted in that direction. We start by looking at the challenges and progress in making larg...

Useful Fields:

This paper addresses a highly relevant issue in the field of AI, emphasizing social responsibility in technology development. The integration of diversity and inclusion into AI design is timely, considering the growing societal concerns about bias and representation. The mention of various projects showcases practical applications and encourages interdisciplinary collaboration, enhancing its overall impact. Moreover, the focus on transparency and trust in AI algorithms speaks to current discussions in AI ethics, making the paper both novel and rigorous.

NL-COMM: Enhanced Video Streaming via Advanced Non-Linear Processing

With video streaming now accounting for the majority of internet traffic, wireless networks face increasing demands, especially in densely populated areas where limited spectral resources are shared a...

Useful Fields:

This article presents a novel approach to optimize video streaming in wireless networks using an advanced non-linear MIMO processing framework, NL-COMM. Its significance lies in its practical demonstration with off-the-shelf user equipment in a fully compliant 3GPP environment. The results of improved stream quality and reduced antenna requirements indicate high potential for enhancing spectral efficiency, addressing the pressing need for better performance in densely populated areas. The robustness of the methodology and its empirical comparisons elevate its relevance.

AdaFV: Accelerating VLMs with Self-Adaptive Cross-Modality Attention Mixture

The success of VLMs often relies on the dynamic high-resolution schema that adaptively augments the input images to multiple crops, so that the details of the images can be retained. However, such app...

Useful Fields:

The article introduces a novel mechanism (self-adaptive cross-modality attention mixture) that enhances the efficiency of Vision-Language Models (VLMs), addressing a significant issue of redundant tokens and potentially improving model performance without additional training costs. Its empirical support from extensive experiments highlights robustness and applicability, making it impactful for future research on VLMs and beyond.

MOGNET: A Mux-residual quantized Network leveraging Online-Generated weights

This paper presents a compact model architecture called MOGNET, compatible with a resource-limited hardware. MOGNET uses a streamlined Convolutional factorization block based on a combination of 2 poi...

Useful Fields:

MOGNET presents a significant advancement in model architecture for resource-limited hardware, with innovative elements such as the use of Cellular Automata for online weight generation and novel training methods for quantization. Its focus on compactness and efficiency, coupled with improved accuracy, makes it highly relevant for fields focused on deploying machine learning in constrained environments.

Make yourself comfortable: Nudging urban heat and noise mitigation with smartwatch-based Just-in-time Adaptive Interventions (JITAI)

Humans can play a more active role in improving their comfort in the built environment if given the right information at the right place and time. This paper outlines the use of Just-in-Time Adaptive ...

Useful Fields:

This article presents a novel approach to enhancing personal comfort in urban environments by leveraging technology and user feedback through Just-in-Time Adaptive Interventions (JITAI). The methodological rigor demonstrated through extensive data collection and participant engagement strengthens the implications of the findings. The intervention's focus on customizable strategies based on individual traits and environmental conditions adds significant value, suggesting practical applications and future research directions in urban design and behavioral interventions.

Synergistic Effects of Natural Biosurfactant and Metal Oxides Modification on PVDF Nanofiber Filters for Efficient Microplastic and Oil Removal

The removal of microplastics and oil from oil-water emulsions presents significant challenges in membrane technology due to issues with low permeability, rejection rates, and membrane fouling. This st...

Useful Fields:

The article presents a novel approach to enhancing PVDF nanofiber membranes which are critical for addressing the growing environmental issue of microplastic and oil contamination in wastewater. The methodological rigor is demonstrated through the combination of materials and treatments resulting in substantial improvements in filtration efficiency and antifouling properties. The findings have practical implications that can advance membrane technology significantly, particularly in environmental remediation and wastewater treatment applications.

Comprehensive Survey of QML: From Data Analysis to Algorithmic Advancements

Quantum Machine Learning represents a paradigm shift at the intersection of Quantum Computing and Machine Learning, leveraging quantum phenomena such as superposition, entanglement, and quantum parall...

Useful Fields:

This article offers a thorough and detailed survey of Quantum Machine Learning (QML), which is a rapidly evolving and highly interdisciplinary field. The comprehensive coverage of both algorithmic advancements and application areas such as healthcare and finance demonstrates its potential for widespread impact. Additionally, the identification of challenges and emerging solutions showcases a clear understanding of the current landscape, making it a valuable resource for researchers and practitioners alike. The novelty and rigor in presenting foundational concepts alongside practical applications boost its relevance significantly.

Confidence Estimation for Error Detection in Text-to-SQL Systems

Text-to-SQL enables users to interact with databases through natural language, simplifying the retrieval and synthesis of information. Despite the success of large language models (LLMs) in converting...

Useful Fields:

The study addresses significant challenges in the field of Text-to-SQL, specifically focusing on confidence estimation and error detection. The use of selective classifiers introduces novel methodologies that could improve the robustness of these systems. Moreover, the empirical results showcasing the performance of various models (T5, GPT-4, Llama 3) provide valuable insights into advancements in model calibration, which is critical for broader adoption of Text-to-SQL technology. The potential implications for improving machine learning models' interpretability and reliability make this research impactful.

Populations of Neutron Star Ultraluminous X-ray Sources: Mind your b's and B's

Ultraluminous X-ray sources (ULXs) with neutron star (NS) accretors challenge traditional accretion models, and have sparked a debate regarding the role of geometrical beaming and strong magnetic fiel...

Useful Fields:

This article presents a novel investigation into the effects of strong magnetic fields on neutron star ultraluminous X-ray sources, challenging traditional models and improving the understanding of these astrophysical phenomena. The methodological rigor in population synthesis modeling enhances its credibility and applicability. The implications for observational astronomy and the deeper understanding of accretion processes make this article highly relevant for future research.

Class Incremental Fault Diagnosis under Limited Fault Data via Supervised Contrastive Knowledge Distillation

Class-incremental fault diagnosis requires a model to adapt to new fault classes while retaining previous knowledge. However, limited research exists for imbalanced and long-tailed data. Extracting di...

Useful Fields:

The article presents a novel approach to an important problem in fault diagnosis, particularly addressing the challenges of learning from imbalanced and limited data, which is highly relevant in many industrial applications. The introduction of a supervised contrastive knowledge distillation method for better representation learning offers a fresh angle, and the detailed experimentation adds methodological rigor and practical relevance.

H-infinity and Complex Interpolation

This note is an (exact) copy of the report of Jaak Peetre, "H-infinity and Complex Interpolation". Published as Technical Report, Lund (1981). Some more recent general references have been a...

Useful Fields:

The article appears to be largely a re-publication of an older report with minor updates. While it may hold historical significance and could be of interest to those studying the evolution of mathematical concepts in H-infinity theory, it lacks new research findings or novel contributions. This limits its impact on advancing current understanding in the field.

Rates of (T-)asymptotic regularity of the generalized Krasnoselskii-Mann-type iteration

In this paper we use proof mining methods to compute rates of ( $T$ -)asymptotic regularity of the generalized Krasnoselskii-Mann-type iteration associated to a nonexpansive mapping $T:X\to ...

Useful Fields:

The article introduces a methodology using proof mining to address asymptotic regularity in fixed-point iterations, which is significant in functional analysis and optimization. The innovative application of proof mining presents a unique approach to quantifying rates, which could be useful for both theoretical and applied contexts.

Merging Models on the Fly Without Retraining: A Sequential Approach to Scalable Continual Model Merging

Deep model merging represents an emerging research direction that combines multiple fine-tuned models to harness their specialized capabilities across different tasks and domains. Current model mergin...

Useful Fields:

The article presents a novel approach for continual model merging that addresses significant limitations in existing methods. The focus on sequential model integration is particularly relevant in practical applications where models need to be updated continuously without retraining. The use of orthogonal projections and adaptive scaling mechanisms showcases methodological rigor and innovation. The empirical results highlight a meaningful accuracy improvement, adding to its impact. However, the level of novelty may be tempered by the foundational concept of weight interpolation, which has been explored before, albeit not in the sequential context proposed here.

Augmenting a Large Language Model with a Combination of Text and Visual Data for Conversational Visualization of Global Geospatial Data

We present a method for augmenting a Large Language Model (LLM) with a combination of text and visual data to enable accurate question answering in visualization of scientific data, making conversatio...

Useful Fields:

This article introduces a novel approach to integrate visual data with large language models, which is a significant advancement in the field of AI and data visualization. The methodology's focus on enhancing LLMs without the need for fine-tuning is especially innovative, addressing a current limitation in LLM applications. Moreover, the ability to use already rendered visualizations broadens its applicability. However, the robustness of results, particularly under varied contexts and datasets, needs further exploration.

RWZC: A Model-Driven Approach for Learning-based Robust Wyner-Ziv Coding

In this paper, a novel learning-based Wyner-Ziv coding framework is considered under a distributed image transmission scenario, where the correlated source is only available at the receiver. Unlike ot...

Useful Fields:

The proposed model-driven Wyner-Ziv coding framework is innovative as it addresses the challenge of non-stationary source correlation, which has been a limitation in conventional approaches. The use of a warping-prediction network demonstrates methodological rigor and suggests robustness in practical applications. The quantified performance improvements in key metrics (PSNR and MS-SSIM) further validate its effectiveness, making it a potentially influential contribution in the field of image transmission and coding.

Multi-task deep-learning for sleep event detection and stage classification

Polysomnographic sleep analysis is the standard clinical method to accurately diagnose and treat sleep disorders. It is an intricate process which involves the manual identification, classification, a...

Useful Fields:

The proposed multi-task deep-learning methodology addresses a critical gap in sleep event detection and classification, enhancing efficiency and accuracy. The integration of advanced object-detection techniques with multi-variate time sequences shows novelty. The rigorous evaluation across multiple datasets strengthens its validity, making this research highly impactful in clinical and computational domains.

On Round Surgery Diagrams For 3-Manifolds

We introduce the notion of round surgery diagrams in $S^3$ for representing 3-manifolds similar to Dehn surgery diagrams. We give a correspondence between a certain class of round surgery diag...

Useful Fields:

This article presents a novel conceptual framework for understanding 3-manifolds through the introduction of round surgery diagrams, which resemble but extend the classic Dehn surgery diagrams. This representation could significantly influence the study of 3-manifolds and knot theory, particularly by establishing a correspondence that allows for a deeper understanding of manifold constructions. The rigorous approach of defining moves akin to those in Kirby calculus supports the robustness of the work, making it a meaningful contribution to the field.

Recovering latent linkage structures and spillover effects with structural breaks in panel data models

This paper introduces a framework to analyze time-varying spillover effects in panel data. We consider panel models where a unit's outcome depends not only on its own characteristics (private effe...

Useful Fields:

This paper presents a novel approach to analyzing panel data models, specifically targeting spillover effects and structural breaks that are crucial for understanding interdependencies in complex systems. Its methodological rigor, particularly the incorporation of penalized estimation and double machine learning, indicates strong innovations that can significantly advance the field of econometrics and social sciences. Furthermore, the practical application to cross-country R&D spillovers highlights its real-world relevance and potential usefulness in policy-making and international collaboration contexts.

Proximal Quasi-Newton Method for Composite Optimization over the Stiefel Manifold

In this paper, we consider the composite optimization problems over the Stiefel manifold. A successful method to solve this class of problems is the proximal gradient method proposed by Chen et al. Mo...

Useful Fields:

The article presents a novel optimization method that addresses composite optimization problems on the Stiefel manifold, which is a significant advancement in the field of Riemannian optimization. The method shows proven global convergence and local linear convergence, indicating robust theoretical underpinnings. The competitive numerical results suggest practical applicability and effectiveness, enhancing the potential influence on further research in optimization techniques. However, the specificity of the application may limit broader interdisciplinary applicability compared to more generalized methods.

On the factorization of twisted $L$ -values and $11$ -descents over $C_5$ -number fields

We investigate the Galois module structure of the Tate-Shafarevich group of elliptic curves. For a Dirichlet character $χ$ , we give an explicit conjecture relating the ideal factorization of &...

Useful Fields:

This article introduces a novel approach to understand the intricate relationship between $L$-values and the Galois module structure of the Tate-Shafarevich group, which is an essential aspect of the arithmetic of elliptic curves. The conjecture presented is explicit and well-supported by numerical evidence and visualization methods, demonstrating strong methodological rigor. Additionally, the focus on practical computation of descents presents substantial applicability, potentially making advanced computational techniques more accessible in this field. However, the niche nature of the topic may limit its immediate impact beyond specific subfields.