Search for: All records

Telescopes such as the Rubin Observatory and Euclid Space telescopes find more and more low mass galaxies as time passes by. However, the link between small dwarf galaxy light and star mass is currently unclear. We recently found out that there are clear differences between stellar mass depending on star formation history (SFH), according to an article published by Mithi A. C. de los Reyes. This is shown in artificial data and we can show this using real observed data. Therefore, using the Galaxy And Mass Assembly (GAMA) survey, we will compare mass estimates for small dwarf galaxies with the utilization of four different methods. We seek to find out if these methods agree. Or do they not agree? What happens when a galaxy is so small it only has a few million suns of stars inside? Are there stars missing? Using jupyter notebooks and python, we can compare the GAMA data and make plots and comparisons to answer these questions. 

Identifying, localizing, and resolving bugs in software engineering is challenging and costly. Approaches to resolve software bugs range from Large Language Model (LLM) code analysis and repair, and automated code repair technology that aims to alleviate the technical burden of difficult to solve bugs. We propose RAGFix, which enhances LLM’s capabilities for bug localization and code repair using Retrieval Augmented Generation (RAG) based on dynamically collected Stack Overflow posts. These posts are searchable via a Question and Answer Knowledge Graph (KGQA). We evaluate our method on the HumanEvalFix benchmark for Python using relevant closed and open-source models. Our approach facilitates error resolution in Python coding problems by creating a searchable, embedded knowledge graph representation of bug and solution information from Stack Overflow, interlinking bugs, and solutions through semi-supervised graph construction methods. We use cosine similarity on embeddings based on LLM-synthesized summaries and algorithmic features describing the coding problem and potential solution to find relevant results that improve LLM in-context performance. Our results indicate that our system enhances small open-source models’ ability to effectively repair code, particularly where these models have less parametric knowledge about relevant coding problems and can leverage nonparametric knowledge to provide accurate, actionable fixes. 

Abstract Digital transformation of manufacturing industry, Smart Manufacturing, leverages continuous measurement of machines on the shop floor to make effective decisions and improve productivity metrics such as machine uptime and overall equipment efficiency (OEE). However, despite the declining sensor cost, the initial financial and technological skill requirements of digital transformation pose significant barriers for the overwhelming majority (90%) of the manufacturers who are classed as small and medium enterprises (SMEs). To lower this barrier, here we demonstrate an inexpensive (~ $40 per machine), data-efficient solution that extracts part-level productivity metrics of a CNC machine from its total current consumption alone. We introduce the concept of a part’s “fingerprint” and develop a set of methods that allows one to extract the fingerprints and utilize them to monitor each individual manufactured part and their cycle times. Testing on actual production data of over 3 three months in a part-counting task, the algorithms show a good match (96.2% overall accuracy) with manually logged production data is achieved. The presented fingerprint framework is general: it can be extended to multi-sensors, and multi-modal analytics. We expect that such a simple, yet cost-effective, solution will be accessible for a wide range of discrete manufacturers, facilitating the beginning of their digital transformation journey. 

Deep neural networks (DNNs) are increasingly used in critical applications like autonomous vehicles and medical diagnosis, where accuracy and reliability are crucial. However, debugging DNNs is challenging and expensive, often leading to unpredictable behavior and performance issues. Identifying and diagnosing bugs in DNNs is difficult due to complex and obscure failure symptoms, which are data-driven and compute-intensive. To address this, we propose TransBug a framework that combines transformer models for feature extraction with deep learning models for classification to detect and diagnose bugs in DNNs. We employ a pre-trained transformer model, which has been trained in programming languages, to extract semantic features from both faulty and correct DNN models. We then use these extracted features in a separate deep-learning model to determine whether the code contains bugs. If a bug is detected, the model further classifies the type of bug. By leveraging the powerful feature extraction capabilities of transformers, we capture relevant characteristics from the code, which are then used by a deep learning model to identify and classify various types of bugs. This combination of transformer-based feature extraction and deep learning classification allows our method to accurately link bug symptoms to their causes, enabling developers to take targeted corrective actions. Empirical results show that the TransBug shows an accuracy of 81% for binary classification and 91% for classifying bug types. 

Protein tandem mass spectrometry (MS/MS) often generates sequence-informative fragments from backbone bond cleavages near the termini. This lack of fragmentation in the protein interior is particularly apparent in native top-down MS. Improved sequence coverage, critical for reliable annotation of posttranslational modifications (PTMs) and sequence variants, may be obtained from internal fragments generated by multiple backbone cleavage events. However, internal fragment assignments can be error prone due to isomeric/isobaric fragments from different parts of a protein sequence. Also, internal fragment generation propensity depends on the chosen MS/MS activation strategy. Here, we examine internal fragment formation in electron capture dissociation (ECD) and electron transfer dissociation (ETD) following native and denaturing MS, as well as liquid chromatography (LC)/MS of several proteins. Experiments were undertaken on multiple instruments, including Q-ToF, Orbitrap, and high-field FT-ICR across four laboratories. ECD was performed at both ultrahigh vacuum and at similar pressure to ETD conditions. Two complementary software packages were used for data analysis. When feasible, ETD-higher-energy collision dissociation (ETD-HCD) MS3 was performed to validate/refute potential internal fragment assignments, including differentiating MS3 fragmentation behavior of radical vs. even-electron primary fragments. We show that, under typical operating conditions, internal fragments cannot be confidently assigned in ECD, nor ETD. On the other hand, such fragments, along with some b-type terminal fragments (not typically observed in ECD/ETD spectra) appear at atypical ECD operating conditions, suggesting they originate from a separate ion-electron activation process. Furthermore, atypical fragment ion types, e.g., x ions, are observed at such conditions as well as upon EThcD, presumably due to vibrational activation of radical z-type ions. 

We revisit the problem of building static hash tables on the GPU and present an efficient implementation of bucketed hash tables. By decoupling the probing scheme from the hash table in-memory representation, we offer an implementation where the number of probes and the bucket size are the only factors limiting performance. Our analysis sweeps through the hash table parameter space for two probing schemes: cuckoo and iceberg hashing. We show that a bucketed cuckoo hash table (BCHT) that uses three hash functions outperforms alternative methods that use iceberg hashing and a cuckoo hash table that uses a bucket size of one. At load factors as high as 0.99, BCHT enjoys an average probe count of 1.43 during insertion. Using three hash functions only, positive and negative queries require at most 1.39 and 2.8 average probes per key, respectively.