Fast Machine Learning for Science Conference 2025

Session

Posters and coffee

2 Sept 2025, 15:00

HIT G floor (gallery)

Presentation materials

All posters

poster_listing_wrapped.pdf

140. A Hybrid Architecture for Real-Time Experiment Feedback in X-ray Microscopy Enabled by Edge and Cloud Computing

Felix Bachmair (Dectris Ltd.)

Posters

Poster

Ptychographic imaging generates high-resolution datasets at the cost of heavy computational complexity, limiting its use in real-time experimental decision-making. In this cross-institutional effort, we introduce a hybrid edge-to-cloud workflow that delivers fast feedback for ptychography experiments by combining a modern synchrotron beamline at Diamond Light Source I13-1, featuring an...

109. Acceleration of a Quantized LeNet-based IDS on FPGA using FINN

Ameth Thiam

Poster

1. Introduction and Context

With the rise of cyberattacks and the growing volume of network traffic, intrusion detection systems (IDS) must provide fast, accurate, and resource-efficient analysis. Traditional CPU- or GPU-based solutions often struggle to meet low-latency and low-power requirements, especially in embedded environments.

Integrating artificial intelligence, particularly...

44. AI-Enhanced Dual-Coated QCM Sensor with FPGA-Based Impedance Analysis for Real-Time Multi-Parameter Environmental Monitoring

N Ramakrishnan (Associate Professor, Monash University Malaysia)

Poster

Quartz Crystal Microbalance (QCM) sensors are renowned for their high sensitivity to mass changes, making them ideal for detecting environmental parameters such as relative humidity (RH) and ultraviolet (UV) radiation. In this work, we present an AI-driven, dual-sided coated QCM sensor integrated with advanced machine learning (ML) and implemented on a real-time hardware platform. This sensor...

137. AutoDeploy-HEP: An Intelligent Toolkit for ML Deployment on Heterogeneous Hardware in High-Energy Physics

MUSTOFA ABDULHAFIZ AHMED mustofa

Poster

Deploying ML models today requires deep expertise in both hardware and software optimization. It often involves laborious trial-and-error to determine the right combination of tools, techniques, and configurations. While industry and academia benefit from a wide array of deployment frameworks and automation tools, the High-Energy Physics (HEP) community still faces major challenges in adopting...

52. Chisel4ml: Direct Circuit Implementation of Deeply Quantized Neural Networks

Jure Vreča

Poster

Chisel4ml is a tool we developed for generating fast implementations of deeply quantized neural networks. The tool has a Python frontend and a Chisel backend. The Python frontend serves as an interface to the Python ecosystem for training neural networks. The Chisel backend consists of hardware generators written in the Chisel Hardware Construction Language. This is a language embedded in...

23. Data-Driven Optimization of Chemical Vapor Deposition Using Machine Learning and Surrogate Modeling

Dr Raja Selvam

Poster

Chemical Vapor Deposition (CVD) optimization is critical for advancing thin-film quality and process efficiency in semiconductor and optoelectronic applications, yet traditional methods like Computational Fluid Dynamics (CFD) simulations and empirical tuning are often computationally intensive or lack adaptability. To address this challenge, this study presents a data-driven machine learning...

104. Dynamic Scheduling Support for Faster ML Inference in hls4ml

Mr Andrei Girjoaba (ETH Zurich)

Poster

FPGAs are performant and flexible microchips well-suited for experimental physics that efficiently run anomaly detection algorithms and identify potential new physical phenomena. However, FPGAs are not easy to program: A significant gap exists between the algorithms used to discover new physics and the low-level hardware description languages (HDLs) required to program FPGAs. To tackle the...

50. Efficient and Interpretable Transformers for Particle Physics

Abhijith Gandrakota (Fermi National Accelerator Lab. (US))

Poster

Transformers excel at modeling correlations in LHC collisions but incur high costs from quadratic attention. We analyze the Particle Transformer using attention maps and pair correlations on the (η,ϕ) plane, revealing that Particle Transformer attention maps learn traditional jet substructure observables. To improve efficiency we benchmark linear attention variants on JetClass and find that...

46. Efficient data movement for Machine Learning inference in heterogeneous CMS software

Davide Valsecchi (ETH Zurich (CH))

Poster

Efficient data processing using machine learning relies on heterogeneous computing approaches, but optimizing input and output data movements remains a challenge. In GPU-based workflows data already resides on GPU memory, but machine learning models requires the input and output data to be provided in specific tensor format, often requiring unnecessary copying outside of the GPU device and...

61. Electron/Positron - Proton Classification with AMS ECAL using Convolutional Vision Transformers

Berk Turk (Middle East Technical University (TR))

Poster

Alpha Magnetic Spectrometer (AMS-02) is a precision high-energy cosmic-ray experiment consisting of Transition Radiation Detector (TRD), Silicon Tracker, Magnet, Time of Flight (ToF), and Ring Imaging Cherenkov Detector (RICH), Anti-Coincidence Counter (ACC), and Electromagnetic Calorimeter (ECAL) on the ISS operating since 2011, and has collected more than 240 billion cosmic-ray events. Among...

77. Enhancing Transit Detection to Uncover Long-Period Exoplanets

David Degen (ETH Zurich, Queloz Group)

Poster

Small ($R<4\,\mathrm{R}_{\oplus}$), long-period ($30\,\mathrm{days}<P$) exoplanets with low equilibrium temperatures are an extremely interesting population, promising insights into planet formation, atmospheric chemistry and evolution, as well as habitability. However, for these planets, the current observing strategy of NASA's Transiting Exoplanet Survey Satellite (TESS) can only capture...

68. EveNet: Towards a Generalist Event Transformer for Unified Understanding and Generation of Collider Data

Yuan-Tang Chou (University of Washington (US))

Poster

With the increasing size of the machine learning (ML) model and vast datasets, the foundation model has transformed how we apply ML to solve real-world problems. Multimodal language models like chatGPT and Llama have expanded their capability to specialized tasks with common pre-train. Similarly, in high-energy physics (HEP), common tasks in the analysis face recurring challenges that demand...

57. Evolution of the oneAPI backend for hls4ml

Jovan Mitrevski (Fermi National Accelerator Lab. (US))

Poster

Since version 1.0, hls4ml has provided a oneAPI backend for Altera FPGAs, as an evolution of the backend that targeted Intel HLS. Some design choices will be presented here, including the use of pipes and task sequences to develop a dataflow-style architecture. The oneAPI framework, unlike the Intel HLS framework, also naturally supports an accelerator-style deployment. Using always-run...

125. Fast Adaptive Neural Control of Resonant Extraction at Fermilab

Maira Khan (Fermi National Accelerator Laboratory)

Poster

We present the development of a machine learning (ML) based regulation system for third-order resonant beam extraction in the Mu2e experiment at Fermilab. Classical and ML-based controllers have been optimized using semi-analytic simulations and evaluated in terms of regulation performance and training efficiency. We compare several controller architectures and discuss the integration of...

74. Fast Synthetic X-Ray Generation for AI Diagnostics: DALL·E vs. Stable Diffusion

RUKSHAK KAPOOR

Poster

Medical imaging is foundational to clinical diagnostics and biomedical research, enabling the identification and monitoring of a wide range of conditions—from pulmonary diseases to cancer. However, the development of high-performance AI diagnostic systems is often hampered by restricted access to large, diverse, and well-annotated imaging datasets. This limitation is particularly acute for...

70. Flexibly Equivariant Generative framework for Stochastic sub-grid Turbulence modeling

Sharvaree Vadgama (University of Amsterdam), Julia Balla (MIT), Ryley McConkey (MIT)

Poster

Introduction
Accurate climate prediction hinges on the ability to resolve multi-scale turbulent dynamics in the atmosphere and oceans [1]. An important mechanism of energy exchange between the ocean and the atmosphere is mesoscale turbulence, which contains motions of length scale $\mathcal{O}$(100 km). Two-layer quasi-geostrophic (QG) simulations [2] are a popular technique for...

55. FloatQuant: Arbitrary-Precision Minifloats in QONNX

Nicolo Ghielmetti (CERN), Yaman Umuroglu (AMD Research)

Poster

The rising popularity of large language models (LLMs) has led to a growing demand for efficient model deployment. In this context, the combination of post-training quantization (PTQ) and low-precision floating-point formats such as FP4, FP6 and FP8 has emerged as an important technique, allowing for rapid and accurate quantization with the ability to capture outlier values in LLMs without...

131. FPGA Toolchain Integration for CI/CD workflows

Andrew Whitbeck (Fermi National Accelerator Lab. (US)), Ben Hawks (Fermi National Accelerator Lab)

Poster

Modern development flows that use tooling for automated building, testing, and deployment of software are becoming the norm for large scale software and hardware projects. These flows offer quite a few advantages that make them desirable, but when attempting to implement them for projects that use FPGAs, some complications can arise when attempting to integrate them with traditional FPGA...

66. FPGA-Optimized ML for Fast electron identification and pT​ regression with the CMS Phase-2 L1 trigger

Piero Viscone (CERN & University of Zurich (CH))

Poster

In preparation for the High Luminosity LHC (HL-LHC) run, the CMS experiment is developing a major upgrade of its Level-1 (L1) Trigger system, which will integrate high-granularity calorimeter data and real-time tracking using FPGA-based processors connected via a high-bandwidth optical network. A central challenge is the identification of electrons in a high pileup environment within strict...

28. Graph Neural Networks for Online Track Reconstruction using FPGAs at the Event Filter for Phase-II Upgrades for the ATLAS Experiment

ATLAS Collaboration

Poster

The High-Luminosity LHC (HL-LHC) will provide an order of magnitude increase in integrated luminosity and enhance the discovery reach for new phenomena. The increased pile-up foreseen during the HL-LHC necessitates major upgrades to the ATLAS detector and trigger. The Phase-II trigger will consist of two levels, a hardware-based Level-0 trigger and an Event Filter (EF) with tracking...

30. High Throughput FPGA Deployment of Distilled Deep Sets Networks for Jet Preselection in the High-Level Trigger

ATLAS TDAQ collaboration, Lucas Bezio (Universite de Geneve (CH))

Poster

Deep Sets-based neural networks are well-suited to learning from unordered, variable-length inputs such as particle tracks associated with jets. Their permutation-invariant structure makes them attractive for high-energy physics (HEP) applications where input ordering is ambiguous and throughput is a critical constraint. In this work, we explore the use of such architectures on...

64. High-Throughput Ghost Track Rejection with Deep Learning at LHCb

Jiahui Zhuo (Univ. of Valencia and CSIC (ES))

Poster

The LHCb experiment at CERN operates a fully software-based first-level trigger that processes 30 million collision events per second, with a data throughput of 4 TB/s. Real-time tracking—reconstructing particle trajectories from raw detector hits—is essential for selecting the most interesting events, but must be performed under tight latency and throughput constraints.
A key bottleneck in...

67. Highly Granular Quantization for CICADA

Abhishikth Mallampalli (University of Wisconsin Madison (US)), Lino Oscar Gerlach (Princeton University (US))

Invited Talks

Poster

The CICADA (Calorimeter Image Convolutional Anomaly Detection Algorithm) project aims to detect anomalous physics signatures without bias from theoretical models in proton-proton collisions at the Compact Muon Solenoid (CMS) experiment at the Large Hadron Collider. CICADA identifies anomalies in low-level calorimeter trigger data using a convolutional autoencoder, whose behavior is...

49. Improving On-Chip Compression of High-Granularity Calorimeter Data with Conditional Autoencoders

Erdem Yigit Ertorer (Carnegie-Mellon University (US))

Poster

The Large Hadron Collider (LHC) will soon undergo a high-luminosity (HL) upgrade to improve future searches for new particles and to measure particle properties with increased precision. The upgrade is expected to provide a dataset ten times larger than the one currently available by the end of its data-taking period. The increased beam intensity will also increase the number of simultaneous...

136. Integrating Support for Google XLS in hls4ml

Andrei Girjoaba

Poster

As Moore’s Law comes to an end, domain-specific architectures (DSA) are considered the next direction for performance improvements in compute. Unfortunately, the development environment of DSAs falls short in comparison to that of general-purpose architectures (e.g., CPUs). The transition from general-purpose to DSA is hindered by the fact that software engineers lack the knowledge to...

138. Inverse Design for Femtosecond-Laser Photonic Surfaces with Direct Gradient Optimization

Dr Amine Haboub (Qatar Environment and Energy Research Institute)

Poster

The inverse design of photonic surfaces produced by high-throughput femtosecond laser processing is limited by a strongly non-linear, many-to-one mapping from laser parameters (power, speed, hatch spacing) to the resulting optical spectrum. Tandem Neural Networks (TNNs) mitigate this ill-posedness by pairing a forward surrogate with a separately trained inverse network, but they still rely on...

110. Locality-Sensitive Hashing-Based Efficient Point Transformer for Charged Particles Reconstruction

Yuan-Tang Chou (University of Washington (US))

Poster

Charge particle track reconstruction is the foundation of the collider experiments. Yet, it's also the most computationally expensive part of the particle reconstruction. The innovation in tracking reconstruction using graph neural networks (GNNs) has demonstrated a promising capability to address the computing challenges posed by the High-Luminosity LHC (HL-LHC) with Machine learning....

94. Loss Landscape Analysis for Reliable Quantized ML Models for Scientific Sensing

Tommaso Baldi (Scuola Superiore Sant'Anna), Dr Tran Nhan (Fermi National Accelerator Laboratory, Batavia, IL, USA)

Poster

In this paper, we propose a method to perform empirical analysis of the loss landscape of machine learning (ML) models. The method is applied to two ML models for scientific sensing, which necessitates quantization to be deployed and are subject to noise and perturbations due to experimental conditions.
Our method allows assessing the robustness of ML models to such effects as a function of...

132. MLCommons Science Benchmarks

Ben Hawks (Fermi National Accelerator Lab)

Poster

Benchmarks are a cornerstone of modern
machine learning practice, providing standardized eval-
uations that enable reproducibility, comparison, and
scientific progress. Yet, as AI systems — particularly
deep learning models — become increasingly dynamic,
traditional static benchmarking approaches are losing
their relevance. Models rapidly evolve in architecture,
scale, and capability;...

122. NomAD: Low-Latency Unsupervised Anomaly Detection for the ATLAS Trigger

Rajat Gupta (University of Pittsburgh (US))

Poster

We present NomAD (Nanosecond Anomaly Detection), a real-time anomaly detection algorithm designed for the ATLAS Level-1 Topological (L1Topo) trigger using unsupervised machine learning. The algorithm combines a Variational Autoencoder (VAE) with Boosted Decision Tree (BDT) regression to compress and distill deep learning inference into a firmware-compatible format for FPGAs. Trained on 2024...

20. Optimizing Reciprocal Human-Machine Learning Decisions

Gila Fruchter

Poster

Recent advances in machine learning have raised ethical concerns in both industry and academia regarding the uncontrollable diffusion of AI and the diminishing human capacity to oversee its impacts. These concerns underscore the need for regulatory and design approaches that maintain human oversight in AI-driven decision-making. Keeping humans in the loop is essential for auditing, fairness,...

34. Planetary systems architecture based on a conditional generative model

Sara Marques (UniBe)

Poster

Understanding the diversity and structure of planetary systems requires capturing not only the properties of individual planets but also the statistical relationships between planets within the same system and their interaction with the host star.

Traditional population synthesis models, such as the Bern model, provide physically motivated insights into these correlations, but their...

95. Porting MADGRAPH to FPGA Using High-Level Synthesis (HLS)

Hector Gutierrez Arance (Univ. of Valencia and CSIC (ES))

Poster

The escalating demand for data processing in particle physics research has spurred the exploration of novel technologies to enhance the efficiency and speed of calculations. This study presents the development of an implementation of MADGRAPH, a widely used tool in particle collision simulations, to FPGA using High-Level Synthesis. This research presents a proof of concept limited to a single,...

157. Poster & Coffee

Poster

114. Real-Time Andean Cryo-Hydrology Intelligence: A Heterogeneous Computing Framework for Glacial Hazard Early Warning

Siwar Jose Basualdo Garcia

Poster

The accelerated retreat of tropical glaciers in the Peruvian Andes poses an imminent and catastrophic threat of Glacial Lake Outburst Floods (GLOF). These events can devastate downstream communities like Huaraz with warning times of less than 15 minutes [1]. Existing monitoring systems are inadequate for this challenge; optical satellite observations (e.g. Landsat, Sentinel-2) are frequently...

45. Real-Time GPU Kalman-Filter Tracking via Kernel Refactoring and INT8 Surrogates for High-Luminosity Colliders

Mr Hao-Chun Liang (Institute of Pioneer Semiconductor Innovation, National Yang Ming Chiao Tung University)

Poster

As the era of the High-Luminosity Large Hadron Collider (HL-LHC) approaches, the GPU-accelerated High-Level Trigger (HLT) of the CMS experiment faces a stringent requirement to reduce the Level-1 readout stream from 100 kHz to 5 kHz, a twenty-fold decrease essential to adhere to archival bandwidth constraints [[1][1]], [[2][2]]. Meeting this demand necessitates highly efficient real-time...

89. Scientific Machine Learning for Symbolic Recovery of Relativistic Effects in Black Hole Orbits

Pothuraju Naveen Yadav (Delhi Technological University)

Poster

Simulating relativistic orbital dynamics around Schwarzschild black holes is essential for understanding general relativity and astrophysical phenomena like precession. Traditional numerical solvers face difficulty while dealing with noisy or sparse data, necessitating data-driven approaches. We develop a Scientific Machine Learning (SciML) framework to model orbital trajectories and...

124. State Space Models for Scientific Time Series Applications

Christina Reissel (Massachusetts Inst. of Technology (US)), Maira Khan (Fermi National Accelerator Laboratory)

Poster

We investigate the application of state space models (SSMs) to a diverse set of scientific time series tasks. In particular, we benchmark the performance of SSMs against a set of baseline neural networks across three domains: magnet quench prediction, gravitational wave signal classification (LIGO), and neural phase estimation. Our analysis evaluates both computational efficiency—quantified by...

106. TDSCAN : Trigger Distributed Spatial Convolution Area Network

Tanguy Dietrich

Poster

Cherenkov Telescope cameras stream about 1 Billion frames per seconds and are dominated by night-sky background, yet the γ-ray air-shower patterns of interest appear only occasionally.
Filtering is thus paramount for guaranteeing that science-grade data are recorded without saturating the downstream read-out.
In this work we present TDSCAN (Trigger Distributed Spatial Convolution Area...

91. TinyML-Based Early Detection of Pak Choi Leaf Diseases Using FPGA in Vertical Farming Environments

PURABI MAZUMDAR. (Centre for Research in Biotechnology for Agriculture, Universiti Malaya, Kuala Lumpur, Malaysia)

Poster

Pak choi (Brassica rapa subsp. chinensis) is a leafy green vegetable widely cultivated in vertical urban farming systems due to its rapid growth and high yield under compact, hydroponic setups. However, even in these controlled environments, crops remain susceptible to various diseases. Among the most common threats are fungal infections such as Alternaria leaf spot and powdery mildew, and...

40. Towards Fast and Interpretable Physics-Informed Learning: Second-Order Neurons and Mixed-Activation Networks

João Paulo De Souza Böger

Poster

Complex simulators are central to scientific research, forecasting, and real-world applications. However, they often require intensive computational resources and suffer from scalability issues — challenges amplified in the big data era. The APEX project addresses these limitations by designing novel efficient architectures for scientific simulators, exploring inductive biases, causal...

126. Towards Online Machine Learning in DUNE Data Acquisition

Olivia Dalager (Fermilab)

Poster

Processing the large volumes of data produced by liquid argon time projection chamber (LArTPC) experiments presents a significant challenge, especially those at the scale of the Deep Underground Neutrino Experiment (DUNE). This is a particular challenge when aiming to trigger on low-energy neutrinos from core-collapse supernovae, which are typically buried in a high-rate radiological...

26. Tracking for the next ATLAS event filter with GNNs on GPUs

ATLAS Collaboration

Poster

Graph Neural Networks (GNNs) have been in the focus of machine-learning-based track reconstruction for high-energy physics experiments during the last years. Within ATLAS, the GNN4ITk group has investigated this type of algorithm for track reconstruction at the High-Luminosity LHC (HL-LHC) using the future full-silicon Inner Tracker (ITk).

The Event Filter (EF) is part of the ATLAS Trigger...

130. wa-hls4ml: A Benchmark and Surrogate Models for hls4ml Resource and Latency Estimation

Ben Hawks (Fermi National Accelerator Lab)

Poster

As machine learning (ML) is increasingly implemented in hardware to address real-time challenges in scientific applications, the
development of advanced toolchains has significantly reduced the time required to iterate on various designs. These advancements have
solved major obstacles, but also exposed new challenges. For example, processes that were not previously considered bottlenecks,...

39. When Less is More: Optimizing Graph Neural Networks and Knowledge Distillation for Efficient Particle Reconstruction and Identification in LHCb's Next-Generation Calorimeter

Cilicia Uzziel Perez (La Salle, Ramon Llull University (ES))

Poster

Graph Neural Networks (GNNs) have become promising candidates for particle reconstruction and identification in high-energy physics, but their computational complexity makes them challenging to deploy in real-time data processing pipelines. In the next-generation LHCb calorimeter, detector hits—characterized by energy, position, and timing—can be naturally encoded as node features, with...