Category: Presentation

Invited Talk by G. Volpe at Gothenburg Lise Meitner Award 2024 Symposium, 27 September 2024

(Image created by G. Volpe with the assistance of DALL·E 2)
What is a physicist to do in the age of AI?
Giovanni Volpe
Gothenburg Lise Meitner Award 2024 Symposium
Date: 27 September 2024
Time: 15:00-15:30
Place: PJ Salen

In recent years, the rapid growth of artificial intelligence, particularly deep learning, has transformed fields from natural sciences to technology. While deep learning is often viewed as a glorified form of curve fitting, its advancement to multi-layered, deep neural networks has resulted in unprecedented performance improvements, often surprising experts. As AI models grow larger and more complex, many wonder whether AI will eventually take over the world and what role remains for physicists and, more broadly, humans.

A critical, yet underappreciated fact is that these AI systems rely heavily on vast amounts of training data, most of which are generated and annotated by humans. This dependency raises an intriguing issue: what happens when human-generated data is no longer available, or when AI begins to train on AI-generated data? The phenomenon of AI poisoning, where the quality of AI outputs declines due to self-referencing, demonstrates the limitations of current AI models. For example, in image recognition tasks, such as those involving the MNIST dataset, AI tends to gravitate towards ‘safe’ or average outputs, diminishing originality and accuracy.

In this context, the unique role of humans becomes clear. Physicists, with their capacity for originality, deep understanding of physical phenomena, and the ability to exploit fundamental symmetries in nature, bring invaluable perspectives to the development of AI. By incorporating physics-informed training architectures and embracing the human drive for meaning and discovery, we can guide the future of AI in truly innovative directions. The message is clear: physicists must remain original, pursue their passions, and continue searching for the hidden laws that govern the world and society.

Seminar by G. Volpe at ESPCI/Sorbonne, Paris, 26 September 2024

(Image by A. Argun)
Deep Learning for Microscopy
Giovanni Volpe
Date: 26 September 2024
Place: ESPCI/Sorbonne, Paris, France

Video microscopy has a long history of providing insights and breakthroughs for a broad range of disciplines, from physics to biology. Image analysis to extract quantitative information from video microscopy data has traditionally relied on algorithmic approaches, which are often difficult to implement, time consuming, and computationally expensive. Recently, alternative data-driven approaches using deep learning have greatly improved quantitative digital microscopy, potentially offering automatized, accurate, and fast image analysis. However, the combination of deep learning and video microscopy remains underutilized primarily due to the steep learning curve involved in developing custom deep-learning solutions. To overcome this issue, we have introduced a software, DeepTrack 2.1, to design, train and validate deep-learning solutions for digital microscopy.

Presentation by G. Wang at ECIS, Copenhagen, 5 September 2024

Schematic and brightfield image (inset) of the movement of 16μm diameter micromotor under the illumination of linearly polarized 1064nm laser. (Image by G. Wang.)
Light-driven metamachines
Gan Wang, Marcel Rey, Antonio Ciarlo, Mohanmmad Mahdi Shanei, Kunli Xiong, Giuseppe Pesce, Mikael Käll and Giovanni Volpe
Date: 5 September 2024
Time: 15:45-16:00

The incorporation of Moore’s law into integrated circuits has spurred opportunities for downsizing traditional mechanical components. Despite advancements in single on-chip motors using electrical, optical, and magnetic drives under ~100 μm, creating complex machines with multiple units remains challenging. Here, we developed a ~10 μm on-chip micromotor using a method compatible with complementary metal oxide semiconductors (CMOS) process. The meta-surface is embedded with the motor to control the incident laser beam direction, enabling momentum exchange with light for movement. The rotation direction and speed are adjustable through the meta-surface, along with the intensity and polarization of applied light. By combining these motors and controlling the configuration, we create complex machines with a size similar to traditional machines below 50um, such as the rotary motion mode of multiple gear coupled gear trains, and the linear motion mode combined with rack and pinion, and combine the micro metal The mirror is introduced into the machine to realize the self-controlled scanning function of the laser in a fixed area.

Presentation by A. Callegari at SPIE-OTOM, San Diego, 22 August 2024

One exemplar of the HEXBUGS used in the experiment. (Image by the Authors of the manuscript.)
Active Matter Experiments with Toy Robots
Angelo Barona Balda, Aykut Argun, Agnese Callegari, Giovanni Volpe
SPIE-OTOM, San Diego, CA, USA, 18 – 22 August 2024
Date: 22 August 2024
Time: 3:00 PM – 3:15 PM
Place: Conv. Ctr. Room 6D

Active matter is based on concepts of nonequilibrium thermodynamics applied to the most diverse disciplines. Active Brownian particles, unlike their passive counterparts, self-propel and give rise to complex behaviors distinctive of active matter. As the field is relatively recent, active matter still lacks curricular inclusion. Here, we propose macroscopic experiments using Hexbugs, a commercial toy robot, demonstrating effects peculiar of active systems, such as the setting into motion of passive objects via active particles, the sorting of active particles based on their mobility and chirality. Additionally, we provide a demonstration of Casimir-like attraction between planar objects mediated by active particles.

Reference
Angelo Barona Balda, Aykut Argun, Agnese Callegari, Giovanni Volpe, Playing with Active Matter, arXiv: 2209.04168

Poster by A. Callegari at SPIE-OTOM, San Diego, 19 August 2024

Trajectory of a hexagonal cluster formed by a transparent particle (blu circle) and six light-absorbing particles (red circles) in a traveling sinusoidal optical pattern, in the absence of thermal noise. The direction of the motion of the optical pattern is given by the arrow. The trajectory’s duration is 30 s. (Image by A. Bergsten.)
Chiral active molecules formation via non-reciprocal interactions
Agnese Callegari, Niphredil Klint, John Klint, Alfred Bergsten, Alex Lech, and Giovanni Volpe
SPIE-OTOM, San Diego, CA, USA, 18 – 22 August 2024
Date: 19 August 2024
Time: 5:30 PM – 7:00 PM
Place: Conv. Ctr. Exhibit Hall A

In 2019, Schmidt et al. demonstrated light-induced assembly of active colloidal molecules. They used two types of colloidal particles in a water-lutidine mixture: one transparent and one slightly absorbing light. In their experiment, this determined a non-reciprocal interaction between light-absorbing and transparent particles and promoted active molecule formation controlled by light. Beyond experimental details, we here explore the effects of this non-reciprocal interaction solely, showing its role in active molecule formation and self-propulsion. Simulation allows for the study of complex light profiles, enabling precise control over assembly and propulsion properties, relevant for targeted microscopic delivery.

Poster by A. Callegari at SPIE-OTOM, San Diego, 19 August 2024

Simplified sketch of the neural network used for the simulations of intracavity optical trapping. (Image by A. Callegari.)
Neural networks for intracavity optical trapping
Agnese Callegari, Mathias Samuelsson, Antonio Ciarlo, Giuseppe Pesce, David Bronte Ciriza, Alessandro Magazzù, Onofrio M. Maragò, Antonio Sasso, and Giovanni Volpe
SPIE-OTOM, San Diego, CA, USA, 18 – 22 August 2024
Date: 19 August 2024
Time: 5:30 PM – 7:00 PM
Place: Conv. Ctr. Exhibit Hall A

Intracavity optical tweezers have been proven successful for trapping microscopic particles at very low average power intensity – much lower than the one in standard optical tweezers. This feature makes them particularly promising for the study of biological samples. The modelling of such systems, though, requires time-consuming numerical simulations that affect its usability and predictive power. With the help of machine learning, we can overcome the numerical bottleneck – the calculation of optical forces, torques, and losses – and reproduce, in simulation, the results in the literature and generalize to the case of counterpropagating-beams intracavity optical trapping.

Poster by A. Callegari at SPIE-OTOM, San Diego, 19 August 2024

Schematic of the scattering of a light ray on a Janus particle. (Image by A. Callegari.)
Janus particles in a travelling optical landscape
Agnese Callegari, Giovanni Volpe
SPIE-OTOM, San Diego, CA, USA, 18 – 22 August 2024
Date: 19 August 2024
Time: 5:30 PM – 7:00 PM
Place: Conv. Ctr. Exhibit Hall A

Janus particles possess dual properties that makes them very versatile for soft and active matter applications. Modeling their interaction with light, including optical force and torque, presents challenges. We present here a model of spherical, metal-coated Janus particles in the geometric optics approximation. Via an extension of the Optical Tweezers Geometrical Optics (OTGO) toolbox, we calculate optical forces, torques, and absorption. Through numerical simulation, we demonstrate control over Janus particle dynamics in traveling-wave optical landscapes by adjusting speed and periodicity.

Poster by M. Granfors at SPIE-ETAI, San Diego, 19 August 2024

GAUDI’s latent space representation of Watts–Strogatz Small-World Graphs. (Image by M. Granfors.)
Global graph features unveiled by unsupervised geometric deep learning
Mirja Granfors, Jesús Pineda, Blanca Zufiria Gerbolés, Jiawei Sun, Joana B. Pereira, Carlo Manzo, and Giovanni Volpe
Date: 19 August 2024
Time: 17:30-19:00 (PDT)

Graphs are used to model complex relationships in various domains, such as interacting particles or neural connections within a brain. Efficient analysis and classification of graphs pose significant challenges due to their inherent structural complexity and variability. Here, an approach is presented to address these challenges through the development of the graph autoencoder GAUDI. GAUDI effectively summarizes graph structures while preserving important topological details through multiple hierarchical pooling steps. This enables the extraction of physical parameters describing the graphs. We demonstrate the performance of GAUDI across diverse graph data originating from complicated systems, including the classification of protein assembly structures from single-molecule localization microscopy data, as well as the analysis of collective behavior and correlations between brain connections and age. This approach holds great promise for examining diverse systems, enhancing our comprehension of various forms of graph data.

Presentation by G. Wang at SPIE-MNM, San Diego, 19 August 2024

Schematic and brightfield image (inset) of the movement of 16μm diameter micromotor under the illumination of linearly polarized 1064nm laser. (Image by G. Wang.)
Light-driven metamachines
Gan Wang, Marcel Rey, Antonio Ciarlo, Mohanmmad Mahdi Shanei, Kunli Xiong, Giuseppe Pesce, Mikael Käll and Giovanni Volpe
Date: 19 August 2024
Time: 16:25-16:40 (PDT)

The incorporation of Moore’s law into integrated circuits has spurred opportunities for downsizing traditional mechanical components. Despite advancements in single on-chip motors using electrical, optical, and magnetic drives under ~100 μm, creating complex machines with multiple units remains challenging. Here, we developed a ~10 μm on-chip micromotor using a method compatible with complementary metal oxide semiconductors (CMOS) process. The meta-surface is embedded with the motor to control the incident laser beam direction, enabling momentum exchange with light for movement. The rotation direction and speed are adjustable through the meta-surface, along with the intensity and polarization of applied light. By combining these motors and controlling the configuration, we create complex machines with a size similar to traditional machines below 50um, such as the rotary motion mode of multiple gear coupled gear trains, and the linear motion mode combined with rack and pinion, and combine the micro metal The mirror is introduced into the machine to realize the self-controlled scanning function of the laser in a fixed area.

Presentation by M. Selin at SPIE-ETAI, San Diego, 19 August 2024

3d Visualization of the full Minitweezers 2.0 system. (Illustration by M. Selin.)
Integrating real-time deep learning for automation of optical tweezers experiments
Martin Selin
SPIE-ETAI, San Diego, CA, USA, 18 – 22 August 2024
Date: 19 August 2024
Time: 4:10 PM – 4:25 PM
Place: Conv. Ctr. Room 6D

The perhaps most widely used tool for measuring forces and manipulating particles at the micro and nano-scale are optical tweezers which have given them widespread adoption in physics, chemistry and biology. Despite advancements in computer interaction driven by large-scale generative AI models, experimental sciences—and optical tweezers in particular—remain predominantly manual and knowledge-intensive, owing to the specificity of methods and instruments. Here, we demonstrate how integrating the components of optical tweezers—laser, motor, microfluidics, and camera—into a single software simplifies otherwise challenging experiments by enabling automation through the integration of real-time analysis with deep learning. We highlight this through a DNA pulling experiment, showcasing automated single molecule force spectroscopy and intelligent bond detection, and an investigation into core-shell particle behavior under varying pH and salinity, where deep learning compensates for experimental drift. We conclude that automating experimental procedures increases reliability and throughput, while also opening up the possibility for new types of experiments.