Alzheimer’s Disease Sequencing Project (ADSP)

 The Alzheimer’s Disease Sequencing Project (ADSP) is a comprehensive, multi-phase national consortium aimed at understanding the genetic basis of Alzheimer’s disease and related dementias. Here are the key aspects of the ADSP:

https://www.nia.nih.gov/research/dn/alzheimers-disease-sequencing-project-consortia

NO gene expression?!

## Genomic Data

- The ADSP involves whole-genome sequencing (WGS) and whole exome sequencing (WES) of samples from various cohorts.

  - **Discovery Phase**: Includes WGS for 584 samples from 113 multiplex families, WES for 5,096 AD cases and 4,965 controls, and WES of an enriched sample set comprising 853 AD cases from multiply affected families and 171 Hispanic controls[2][5][6].

  - **Follow-Up Study Phases**: The project has progressed through several phases, including the Discovery Extension Phase, Follow-Up Study Phase, and Follow-Up Study 2.0 Diversity Initiative Phase, which focus on expanding the genetic data to include more diverse populations, such as African Americans, Hispanics, and Asians[1][5].

## Phenotypic Data

- While the primary focus of the ADSP is on genomic data, it also incorporates rich phenotypic data.

  - **Clinical and Cognitive Data**: The project includes clinical cognitive data such as memory, language, and executive function scores. However, it does not directly collect neuroimaging data like T1 MRI, Amyloid-beta, or tau PET scans as part of its core sequencing efforts. Instead, these data are often integrated from other studies and consortia[1][3][6].

  - **Longitudinal and Autopsy-Confirmed Data**: The project emphasizes the use of well-phenotyped participants with autopsy-confirmed diagnoses and longitudinal data[2][5].

## Harmonized Data

- The ADSP Phenotype Harmonization Consortium (ADSP-PHC) plays a crucial role in harmonizing phenotypic data across different cohorts.

  - **ADSP-PHC**: Established to harmonize endophenotype data, including cognitive, imaging, longitudinal clinical, neuropathological, cardiovascular risk, and biomarker data. This harmonization enables modern genomic analyses and generates a perpetually curated and shared legacy dataset[3][6].

## Study Design and Objectives

- The ADSP uses both case-control and family-based study designs.

  - **Objectives**: The overarching goals include identifying new genes involved in Alzheimer’s disease, identifying gene alleles contributing to increased risk or protection against the disease, understanding why individuals with known risk factor genes do not develop AD, and identifying potential therapeutic approaches and prevention strategies[1][4][5].

## Diversity and Global Collaboration

- The ADSP places a high priority on racial/ethnic diversity, recognizing that most genetic studies have been conducted in non-Hispanic white populations.

  - **Diverse Population Initiative**: The Follow-Up Study 2.0 phase aims to conduct whole-genome sequencing on 18,500 AD cases and 18,500 controls from African American, Hispanic, and Asian populations, ensuring a more diverse sample set[1][2][5].

The ADSP is a collaborative effort involving over 350 investigators from global institutions, funded under several cooperative agreements and research grant awards, and is part of the NIA Alzheimer’s Disease Genetics Portfolio.

Citations:

[1] https://www.nia.nih.gov/research/dn/alzheimers-disease-sequencing-project-consortia

[2] https://dss.niagads.org/studies/sa000001/

[3] https://www.vumc.org/cnt/harmonization-initiative

[4] https://www.ncbi.nlm.nih.gov/projects/gap/cgi-bin/study.cgi?study_id=phs000572.v1.p1

[5] https://adsp.niagads.org/about/adsp-phases/

[6] https://adsp.niagads.org/funded-programs/phenotype-harmonization/

[7] https://alz-journals.onlinelibrary.wiley.com/doi/10.1002/alz.13705

[8] https://adsp.niagads.org/adsp-and-affiliates-whole-genome-sequencing-report/

free tools that can detect AI-generated content

 Here are some free tools that can detect AI-generated content with a limit of up to 5,000 words:

1. **SEO.ai**:

   - **Word Limit**: 5,000 characters.

   - **Features**: Provides a probability score indicating whether the content is AI-generated. It uses an ensemble model for more stable results, making it a reliable choice for detecting AI content[2].

2. **Smodin**:

   - **Word Limit**: No specific character limit mentioned, but it supports document uploads (PDF, DOC, DOCX).

   - **Features**: Offers a simple interface for pasting text or uploading documents. It can handle multiple languages and provides a probability score for AI detection[1][3].

3. **QuillBot**:

   - **Word Limit**: No strict limit mentioned; however, it typically processes smaller texts effectively.

   - **Features**: Offers an overall percentage likelihood of AI generation and categorizes the text into different classifications (AI-generated, human-written, etc.). It does not require sign-up for use[1][4].

4. **GPTZero**:

   - **Word Limit**: Up to 5,000 characters.

   - **Features**: Allows users to input text directly or upload documents. It analyzes the text quickly and provides insights on whether the content is likely human or AI-generated[3][7].

5. **Leap AI**:

   - **Word Limit**: Not specified, but allows document uploads.

   - **Features**: Provides a percentage score estimating AI involvement and highlights sentences with high scores for AI generation[1].

These tools vary in their specific capabilities and user interfaces, but they all provide free options for detecting potential AI-generated content effectively.

Citations:

[1] https://surferseo.com/blog/best-ai-content-detection-tools/

[2] https://seo.ai/blog/free-ai-content-detectors

[3] https://zapier.com/blog/ai-content-detector/

[4] https://www.scribbr.com/ai-tools/best-ai-detector/

[5] https://originality.ai/blog/best-ai-content-detection-tools-reviewed

[6] https://contentdetector.ai

[7] https://www.twixify.com/post/best-ai-content-detectors

EU software legislations

 

What the EU’s new software legislation means for developers

https://github.blog/open-source/maintainers/what-the-eus-new-software-legislation-means-for-developers/

Everything you never wanted to know about the R vulnerability ...but shouldn't be afraid to ask

 

Everything you never wanted to know about the R vulnerability

...but shouldn't be afraid to ask

https://aitap.github.io/2024/05/02/unserialize.html

R-bitrary Code Execution: Vulnerability in R’s Deserialization

  R-bitrary Code Execution: Vulnerability in R’s Deserialization

https://hiddenlayer.com/innovation-hub/r-bitrary-code-execution/

NIST National Vulnerability Database

 

https://nvd.nist.gov/

NIST National Vulnerability Database

CWE-502: Deserialization of Untrusted Data

CWE-502: Deserialization of Untrusted Data

https://cwe.mitre.org/data/definitions/502.html

New R programming vulnerability exposes projects to supply chain attacks:

 

https://thehackernews.com/2024/04/new-r-programming-vulnerability-exposes.html

How about citing the following R risk issue before R.4.3.1. 

 

New R programming vulnerability exposes projects to supply chain attacks: 

https://thehackernews.com/2024/04/new-r-programming-vulnerability-exposes.html

A critical security vulnerability, CVE-2024-27322, has been identified in R versions 1.4.0 through 4.3.1. This flaw allows attackers to execute arbitrary code by exploiting the deserialization process of untrusted data, particularly through maliciously crafted RDS (R Data Serialization) files or R packages. The issue stems from R's handling of promise objects and lazy evaluation, enabling an attacker to embed arbitrary R code within an RDS file that executes upon loading and accessing the associated object. This vulnerability poses significant risks in environments where R packages are shared, potentially leading to widespread supply chain attacks. 

 

This issue was fixed in R4.4.0.

 

https://journals.plos.org/ploscompbiol/article?id=10.1371%2Fjournal.pcbi.1012469&utm_source=chatgpt.com#sec008

https://doi.org/10.1371/journal.pcbi.1012469

• 
  

Genome-wide association study between SARS-CoV-2 single nucleotide polymorphisms and virus copies during infections

• Ke Li ,
• Chrispin Chaguza,
• Julian Stamp,
• Yi Ting Chew,
• Nicholas F. G. Chen,
• David Ferguson,
• Sameer Pandya,
• Nick Kerantzas,
• Wade Schulz,
• Yale SARS-CoV-2 Genomic Surveillance Initiative ,
• Anne M. Hahn,
• C. Brandon Ogbunugafor,
• Virginia E. Pitzer,
•  [ ... ],
• Nathan D. Grubaugh 
• [ view all ]

•                     Version 2                              Version 1     Uncorrected Proof               
• Published: September 17, 2024
• https://doi.org/10.1371/journal.pcbi.1012469

Attention-based fusion methods

 Attention-based fusion methods are sophisticated techniques used to combine information from multiple modalities or features, and they do not necessarily require token and vocabulary matching in the traditional sense. Here’s a detailed explanation based on the provided sources:

## Attention Mechanism

The core of attention-based fusion is the attention mechanism, which dynamically adjusts the relative importance of different modalities or features based on the context. This is achieved by computing attention weights that reflect how relevant each modality or feature is to the current task or state.

## Multimodal Attention Fusion

In the context of multimodal fusion, such as in video description or vision-language tasks, attention-based methods allow the model to selectively focus on different modalities (e.g., image, audio, text) when generating outputs. For example:

- The method proposed by Hori et al. uses an attention model to handle the fusion of multiple modalities, where each modality has its own sequence of feature vectors. The attention weights are computed based on the decoder state and the feature vectors, allowing the model to dynamically adjust the importance of each modality[1].

## Channel Fusion and Compound Tokens

In vision-language tasks, methods like Compound Tokens fusion use cross-attention to align visual and text tokens. Here, the model does not require exact token matching but instead uses cross-attention to retrieve compatible tokens from different modalities. The visual and text tokens are then concatenated along the channel dimension to form compound tokens, which are fed into a transformer encoder. This approach does not necessitate a direct match between tokens but rather aligns them through cross-attention[2].

## Attentional Feature Fusion

For feature fusion within neural networks, attention-based methods can be applied across different layers and scales. For instance, the Attentional Feature Fusion (AFF) framework generalizes attention-based feature fusion to cross-layer scenarios, including short and long skip connections. This method uses multi-scale channel attention to address issues arising from feature inconsistency across different scales, without requiring token or vocabulary matching[3].

## Multi-criteria Token Fusion

In the context of vision transformers, Multi-criteria Token Fusion (MCTF) optimizes token fusion by considering multiple criteria such as similarity, informativeness, and size. This method uses one-step-ahead attention to measure the informativeness of tokens and does not require a direct match between tokens. Instead, it aggregates tokens based on their relevance and informativeness, minimizing information loss[4].

## Conclusion

Attention-based fusion methods are highly flexible and do not require explicit token or vocabulary matching. They dynamically adjust the importance of different modalities or features based on the context, allowing for more effective and adaptive fusion of information. These methods are applicable across various domains, including multimodal fusion, vision-language tasks, and feature fusion within neural networks.

Citations:

[1] https://openaccess.thecvf.com/content_ICCV_2017/papers/Hori_Attention-Based_Multimodal_Fusion_ICCV_2017_paper.pdf

[2] https://openreview.net/pdf?id=J9Z3MlnPU_f

[3] https://openaccess.thecvf.com/content/WACV2021/papers/Dai_Attentional_Feature_Fusion_WACV_2021_paper.pdf

[4] https://openaccess.thecvf.com/content/CVPR2024/papers/Lee_Multi-criteria_Token_Fusion_with_One-step-ahead_Attention_for_Efficient_Vision_Transformers_CVPR_2024_paper.pdf

[5] https://www.nature.com/articles/s41598-023-50408-6

[6] https://pmc.ncbi.nlm.nih.gov/articles/PMC9462790/

[7] https://openaccess.thecvf.com/content/CVPR2024/papers/Marcos-Manchon_Open-Vocabulary_Attention_Maps_with_Token_Optimization_for_Semantic_Segmentation_in_CVPR_2024_paper.pdf

newer AI techniques and trends that have emerged or gained significant traction after 2023:

 Here are some of the newer AI techniques and trends that have emerged or gained significant traction after 2023:

## Multimodal AI

Multimodal AI combines multiple modalities such as text, images, audio, and video to create more versatile and effective AI models. This approach allows models like ChatGPT-4 to generate text from various inputs, including images and audio, and to convert between different modalities seamlessly. This trend is expected to enhance applications in fields like financial services, customer analytics, and marketing[1][5][7].

## Small Language Models (SLMs)

SLMs are smaller versions of large language models (LLMs) that can operate efficiently on fewer computing resources, making them accessible on devices like smartphones. These models, such as Microsoft's Phi and Orca, offer similar or sometimes better performance than LLMs in certain areas, democratizing AI use and reducing the need for significant financial investments[1].

## Customizable Generative AI

Customizable AI models are designed to cater to specific industries and user needs, offering more personalization and control over data. This is particularly beneficial in sectors like healthcare, legal, and financial services, where specialized terminology and practices are crucial. Customizable models also enhance privacy and security by reducing reliance on third-party data processing[1].

## Decoupled Contrastive Learning (DCL)

DCL is a new approach to contrastive learning that improves learning efficiency by removing the negative-positive-coupling (NPC) effect present in traditional contrastive learning methods like InfoNCE. This method requires fewer computational resources, smaller batch sizes, and shorter training epochs, yet achieves competitive performance with state-of-the-art models[2][4].

## Explainable AI (XAI)

XAI focuses on making AI models more transparent and interpretable by providing insights into how the models arrive at their decisions. Techniques such as decision trees, linear models, and rule-based systems are used to ensure that AI-driven decisions align with human values and expectations. This trend is gaining popularity as it builds trust and understanding in AI-generated outcomes[3].

## Agentic AI

Agentic AI represents a shift from reactive to proactive AI systems. These AI agents exhibit autonomy, proactivity, and the ability to act independently, setting goals and taking actions without direct human intervention. Applications include environmental monitoring, financial portfolio management, and other areas where autonomous decision-making is beneficial[7].

## Multi-view Graph Contrastive Learning

This approach adapts contrastive learning to recommendation systems by incorporating multiple views of user data. Techniques such as node dropout, edge dropout, and random walks are used to generate diverse views, enhancing the model's ability to capture underlying preferences and behaviors[6].

## Language-Image Contrastive Learning with Efficient Large Language Model and Prompt Fine-Tuning (CLEFT)

CLEFT is a novel method that combines efficient large language models with prompt fine-tuning for language-image contrastive learning. This approach reduces the need for extensive GPU resources and prolonged training times, making it suitable for applications with limited datasets, such as medical imaging[9].

## Retrieval-Augmented Generation

This trend involves combining generative AI models with retrieval systems to enhance the accuracy and relevance of generated content. By retrieving relevant information from a database and integrating it into the generation process, models can produce more informed and contextually accurate outputs[7].

These techniques and trends highlight the rapid evolution and diversification of AI, enabling more efficient, versatile, and interpretable AI applications across various domains.

Citations:

[1] https://khoros.com/blog/ai-trends

[2] https://ai.meta.com/research/publications/decoupled-contrastive-learning/

[3] https://devabit.com/blog/top-11-new-technologies-in-ai-exploring-the-latest-trends/

[4] https://www.amazon.science/blog/new-contrastive-learning-methods-for-better-data-representation

[5] https://www.ibm.com/think/insights/artificial-intelligence-trends

[6] https://www.nature.com/articles/s41598-024-73336-5

[7] https://www.techtarget.com/searchenterpriseai/tip/9-top-AI-and-machine-learning-trends

[8] https://gram-blogposts.github.io/blog/2024/contrast-learning/

[9] https://arxiv.org/abs/2407.21011

AI by Hand

 https://aibyhand.substack.com/t/workbook

AI by Hand