Tero Keski-Valkama

Proud of Owkin K OS latest work: FedPyDESeq2, a privacy-enhancing tool for differential expression analysis in #federatedlearning, is now available on bioRxiv. FedAI breaks research and data silos for an open ecosystem of research. Differential gene expression analysis (DGEA) is one of the most widely used computational techniques to reveal gene expression differences between two or more conditions, a key step to discover new biology and further our understanding of diseases. However, large-scale gene expression datasets, are often stored behind hospitals and institutes' firewalls, limiting opportunities to run large studies across them. Well, no more. FedPyDESeq2, based on Owkin's Substra, allows to perform DGEA in a real-world, federated setting. We rigorously demonstrate that it does so while maintain almost-identical performances to a pooled setting. And yes, it's open-source - because this is what we do Check out the code here: https://lnkd.in/ghSPXARf Congrats to our amazing scientists Boris Muzellec, Ulysse MARTEAU-FEREY and Tanguy Marchand under Eric Durand scientific supervision. Read all about our method here: https://lnkd.in/gFhN679y Importantly, FedPyDESeq2 is built on PyDESeq2, our open-source python implementation of the famous DESeq2 method first introduced by Love, Huber and Anders (2014). #goowkin

55

1 comentario

Taking processing to where the data is is not a trend, it is an obligation, and the future will only be decentralised and data will only be processed in "edge" environments. Today José María Álvarez-Pallete commented in the Wake Up Spain 2024 on the same but referring to Telefónica's networks and we also believe that processing where it will have value is where the non-travelling data is. That is why we are already working on it and one of the Scalian Spain training teams will present a project in C++ of embedded systems with inference of a Machine Learning process in IoT environments with a microcontroller and TinyML at the Universidad Carlos III de Madrid A work that two colleagues from Scalian will make possible live and direct, Silvia Hernández Caballero and Alejandro Hidalgo Paniagua will make a live demo of how to process information in edge environments with low-cost devices, on April 26th. A technological challenge that we will make possible thanks to using std::cpp with J. Daniel Garcia as master of ceremonies and to the fuel it puts in this Formula 1 Susana Santos Carreras. --- "Machine Learning Applications for Embedded Devices Using TinyML and C++ Advancements in the fields of Internet of Things (IoT) and Industry 4.0, for example, require the usage of ML and AI models for several purposes than it would otherwise be impossible to address them. Despite of this, the embedded devices used in almost IoT-based (or Industry 4.0-based) applications have limited computing resources to run the ML or AI models we know so far with agility. Taking this scenario into account, TinyML is a ML tool that has the ability to perform on-device analysis for different sensing modalities such as audio, vision and voice. ML models based on the TinyML tool have low power, memory and computational requirements. These features make it an ideal choice for deploying ML and IA models in the IoT framework. Of course, this optimisation is possible thanks to the C++ programming language, in which this tool is written." https://lnkd.in/emwXEntR

6

【Script-Based GPU-Ready ELM Development for Continuous Code Integration】 Full article: https://lnkd.in/gFM7gFdQ (Authored by Peter Schwartz, et al., from Oak Ridge National Laboratory, Environmental Sciences Division (USA), etc.) The E3SM Land Model (#ELM) is an integral part of the Energy Exascale Earth System Model (E3SM) framework, simulating the interactions between terrestrial land surfaces and other Earth system components. ELM has been instrumental in understanding hydrologic cycles, biogeophysics, and dynamics of terrestrial ecosystems. To facilitate the automatic porting of individual ELM modules, the researchers have developed a #software tool named SPEL. This work presents a SPEL extension that leverages the version control system (e.g., Git) to autonomous code generation and demonstrate its application to #continuous_code_integration and development of the ELM software system.  

2

Three methods will be the foundation for AGI: RGM (Renormalising Generative Models - Friston et al, 2024), JEPA (Joint Embedding Predictive Architecture - LeCun, 2022++), and CORTECONs(R) (COntext-Retentive, TEmporally-CONnected neural networks - Maren, 1991-93, 2014-present). #agi #ai #artificialintelligence #artificialgeneralintelligence #CORTECON #JEPA #RGM #activeinference https://lnkd.in/gDuabFdc

10

This paper investigates the problem of #direction-#of-#arrival (#DOA) estimation using multiple partially-calibrated sparse subarrays. In particular, we present the #Generalized #Coarray #Multiple #Signal #Classification (#GCA-#MUSIC) #DOA #estimation algorithm to scenarios with partially-calibrated sparse subarrays. The proposed #GCA-#MUSIC algorithm exploits the difference coarray for each subarray, followed by a specific pseudo-spectrum merging rule that is based on the intersection of the signal subspaces associated to each subarray. ----@Wesley S. Leite, @Rodrigo C. De Lamare More details can be found at this link: https://lnkd.in/gYBMTKBN

10

This morning, I read an interesting article: "Benchmarking Optimizers for Qumode State Preparation with Variational Quantum Algorithms" by Shuwen Kan et al. Abstract: —Quantum state preparation involves preparing a target state from an initial system, a process integral to applications such as quantum machine learning and solving systems of linear equations. Recently, there has been a growing interest in qumodes due to advancements in the field and their potential applications. However there is a notable gap in the literature specifically addressing this area. This paper aims to bridge this gap by providing performance benchmarks of various optimizers used in state preparation with Variational Quantum Algorithms. We conducted extensive testing across multiple scenarios, including different target states, both ideal and sampling simulations, and varying numbers of basis gate layers. Our evaluations offer insights into the complexity of learning each type of target state and demonstrate that some optimizers perform better than others in this context. Notably, the Powell optimizer was found to be exceptionally robust against sampling errors, making it a preferred choice in scenarios prone to such inaccuracies. Additionally, the Simultaneous Perturbation Stochastic Approximation optimizer was distinguished for its efficiency and ability to handle increased parameter dimensionality effectively. Link: https://lnkd.in/dkB8hBHC #quantumalgorithms #quantummachinelearning #quantumcomputing #quantumoptimizers

30

This insightful document defines quantum machine learning as the integration of a small-scale quantum circuit with a classical optimizer (e.g., SGD, Adam). As such, understanding the fundamental principles and design of quantum computing is a prerequisite. #machinelearning #quantumcomputing #optimizer

7

AI is transforming design—but what about spatial computing? In Bezi Co-founder Cecilia Uhr's #Config2024 talk, we learn more about the tools, challenges, and opportunities in creating immersive 3D experiences. Some key takeaways: ➔ Both AI and 3D tools are advancing rapidly, and designers have the exciting opportunity to start designing for new dimensions, senses, and variables. ➔ Platforms like Bezi empower designers to innovate with interactive experiences, bridging the gap between design and prototyping. ➔ AI and spatial computing offer the potential to create adaptive environments and redefine human interaction, paving the way for personalized experiences. Watch the full talk here: https://lnkd.in/gJTj_84g

8

#mdpisymmetry Check this newly published article "Explicit and Implicit Preference for Symmetry Across Object Categories" at https://brnw.ch/21wOqXR Authors: Marco Roccato et al. Università degli Studi di Padova #symmetry #preference #aesthetics #landscapes

1

Sustainable autoscaling your cloud VM's / Kubernetes with VASIM? —Vertical Autoscaling Simulator Toolkit— Source of paper: https://lnkd.in/e2pYxTSF By Anna Pavlenko, Karla Saur, Yiwen Zhu, Brian Kroth, Joyce Cahoon, Jesús Camacho Rodríguez. Introducing VASIM—Vertical Autoscaling Simulator Toolkit—a cutting-edge tool designed unintendedly to enhance sustainability and efficiency in cloud computing, particularly beneficial for green initiatives in technology (Green IT). Autoscaling, a fundamental aspect of cloud management, can be complex to optimize. VASIM simplifies this process by recommending the best ways to adjust resources based on usage data, ensuring optimal performance while minimizing costs. This approach not only improves efficiency but also supports eco-friendly practices. 🌱 VASIM features an easy-to-use interface that helps evaluate and fine-tune autoscaling strategies using historical data. By simulating scenarios and providing insights into CPU usage and limits, VASIM enables faster development and more efficient resource allocation without needing real-time testing. 💻 One of VASIM's standout benefits is its ability to refine algorithms in a controlled environment, resulting in cost savings and improved resource efficiency. By resolving conflicting autoscaling settings, VASIM ensures that cloud systems operate optimally without unnecessary waste. 📊 This toolkit is adaptable across various workloads, making it accessible and reliable for different applications. Looking forward, VASIM's expansion to include other system components like memory and network performance will provide even deeper insights for informed decision-making and proactive optimization. VASIM is set to transform cloud management by driving efficiency and aligning with sustainable technology practices. 🌍 #greenit #greensoftware #sustainability #softwaredevelopment #k8s #kubernetes #developers #multicloud

7

About the array vs. point receiver debate:  I'm a fan of old analog geophone string arrays.  They clearly work. But with the new cheap nodes (STRYDE, Quantum, NuSeis?, Wing?), it is finally time to retire the old analog string arrays. We can produce equal or better quality data with fewer nodes than with the number of geophones in all the arrays. (e.g. ~60,000 nodes are better than 20,000 array channels with 12 geophones per array, 240,000 geophones total.)   The trick is that you need enough of the nodes and you need good processing technology.  With nodes we can apply statics, SC corrections, and do better velocity filtering. This is good, but we can do even better with optimal phase corrections and this may finally achieve the full potential of nodes in noisy settings.  The problem is that statics, SC corrections, and velocity filtering assume simple noise, but often noise is more complex.  Optimal phase corrections are essentially enhancements to statics, SC corrections, and velocity filtering to account for wave equation effects of a complex near surface. Our trick is figuring out how to do the corrections early in a flow, which then helps all other processing.  I have some 2D tests (see below) that look good. But they are not ideal and I want tests that are larger and more custom! This may finally achieve the full potential of nodes in noisy settings! A key concept is that using digital arrays lets us do coherent noise removal which is more effective than sqrt(N). Can we do an add-on test to your acquisition?  We need regular or irregular node spacing of ~15m or less with moderate shot generated noise.

101

8 comentarios

"NAR-*ICP: Neural Execution of Classical ICP-based Pointcloud Registration Algorithms" by Efimia Panagiotaki, Daniele De Martini, Lars Kunze, and Petar Veličković "This study explores the intersection of neural networks and classical robotics algorithms through the Neural Algorithmic Reasoning (NAR) framework, allowing to train neural networks to effectively reason like classical robotics algorithms by learning to execute them. Algorithms are integral to robotics and safety-critical applications due to their predictable and consistent performance through logical and mathematical principles. In contrast, while neural networks are highly adaptable, handling complex, high-dimensional data and generalising across tasks, they often lack interpretability and transparency in their internal computations. We propose a Graph Neural Network (GNN)-based learning framework, NAR-*ICP, which learns the intermediate algorithmic steps of classical ICP-based pointcloud registration algorithms, and extend the CLRS Algorithmic Reasoning Benchmark with classical robotics perception algorithms. We evaluate our approach across diverse datasets, from real-world to synthetic, demonstrating its flexibility in handling complex and noisy inputs, along with its potential to be used as part of a larger learning system. Our results indicate that our method achieves superior performance across all benchmarks and datasets, consistently surpassing even the algorithms it has been trained on, further demonstrating its ability to generalise beyond the capabilities of traditional algorithms." Paper: https://lnkd.in/dyNahNVT #graphneuralnetworks #machinelearning

39

Hey guys, if you want a really good random number generator for your usecases, do check this open source Library out! Do tag us if you're using and let us know your usecases!

4

This article by Pere Martra explores a structured pruning technique for state-of-the-art models, that uses a GLU architecture, enabling the creation of smaller and more efficient large language models. #Llama #LLM

5

🤖 Quantinuum H2 improvements say a lot about quantum advantage. This is a bit of a different approach to evaluation than previously (though if you want a brief breakdown on the technical implementation, check out today’s newsletter). But, this paper from Quantinuum is significant in its ties to the bigger picture. TLDR: Improvements to Quantinuum's H2 quantum processor allow it to operate with up to 56 qubits and high connectivity while maintaining high two-qubit gate fidelity, as demonstrated with random circuit sampling tasks which are known to be challenging for classical computers. Overall, the improvements to H2 widen the gap between classical ability and quantum ability. As a writer and reader within the space, I can tell you that it’s easy to get lost in the increasingly dense zoos of algorithms, error correction codes, minutiae of various updates to modality architectures, etc. As such, I’ve been considering personal benchmarks for evaluating the value of new papers they can be briefly summed up as follows: ⚆ We need to get answers that weren't accessible before (this can include getting "better" answers), and these answers must be tied to practical applications. ⚆ Energy efficiency must be considered -- if we get a better answer at a significantly higher energy cost, was this truly better than the classically “good” answer? (aka practicality in implementation) ⚆ And, the benefit of applying the solution to the practical application must outweigh the energy cost of achieving that solution. Keeping this in mind, Quantinuum set about to improve their existing hardware in terms of connectivity and fidelity. An aside on connectivity because it doesn't always get the recognition it deserves -- connectivity refers to the capabilities of the hardware in terms of the arrangement of qubits and the available gate operations. Higher connectivity means more direct implementation of quantum gates between a pair of qubits. This helps entanglement as it reduces the number of intermediate operations needed and as such, improves fidelity. Higher connectivity → better entanglement → higher fidelity → better hardware. Plus, better entanglement means more difficulty for classical algorithms to replicate. In terms of previously declared KPIs: ⚆ Advantage over classical solvers is present. ⚆ More efficient hardware implies opportunity for strides towards conservative energy use. Ultimately, focusing on crafting better hardware precedes and leads to finding practical applications. It is these incremental improvements that incrementally increase my already present faith in the technology itself. I look forward to seeing what practical applications the Quantinuum teams' efforts lead to -- congratulations to the team! #quantumcomputing #innovation

6