arXiv subjects

David Steiner

Publications and source records attributed to David Steiner.

8 recordsLinked to original sources

arXiv2026

Accelerating Scientific Research with Gemini in the Real-World

We present an extension and comprehensive real-world validation of Co-Scientist, a Gemini-based multi-agent system designed to accelerate end-to-end scientific research across hypothesis generation, experimentation, and manuscript generation. Moving beyond in silico hypothesis generation, this specialized configuration transitions Co-Scientist into an execution-grounded research partner advancing closed-loop scientific workflows across materials science, biology, and computer science. In materials science, Co-Scientist interfaced with a semi-automated chemical vapor deposition reactor to design a safe precursor route for MXenes; experimental execution produced a lamellar 2D material sharing key structural similarities with the Ti3C2Tx MXene lattice, although further experiments are needed to confirm the atomic structure. Leveraging Gemini 3 Deep Think for rapid, lab-in-the-loop execution, it also tailored growth recipes to laboratory constraints in minutes, enabling single-attempt growth of monolayer MoS2, MoSe2, and WS2 semiconductors. In biology, Co-Scientist predicted emergent swarming phenotypes of engineered E. coli across inducer (IPTG) gradients from sparse imaging data, quantitatively matching unpublished wet-lab morphological measurements. In computer science, Co-Scientist autonomously discovered an inference-time scaling architecture that outperformed six frontier models on HealthBench (Hard and Professional) while reducing potential clinical harm under blinded physician evaluation. Finally, a double-blind study of end-to-end generated papers with 30 domain experts across 450 reviews demonstrates that Co-Scientist's reliability modules reduce hallucination and plagiarism while improving research safety. Together, these results demonstrate progress toward closed-loop multi-agent scientific AI systems capable of accelerating real-world scientific discovery.

cs.AI↗

arXiv2025

Free expansion of a charged nanoparticle via electrostatic compensation

Coherent wavepacket expansion is a key component of recent proposals aiming to create non-classical states of a levitated dielectric nanoparticle. Free evolution, i.e., releasing the particle from its harmonic trapping potential and allowing its position variance to grow, is a simple but effective expansion scheme, but requires accurate force compensation to avoid significant mean displacements of the wavepacket during the free evolution. Here, using an optical trap-release-recapture sequence, we demonstrate an electrostatic compensation technique that enables free evolution without significant mean displacement in 3D, successfully compensating both gravity and stray electric fields. To achieve 100 microsecond free evolution times with charged particles, we developed methods to map and correct for force cross-talk, as well as to control the environmental charge state. Combined with a low decoherence environment, our approach enables the preparation of largely delocalized states without the need for long freefall trajectories or a low-gravity environment.

quant-ph↗

arXiv2024

Tutorial for the growth and development of Myxococcus xanthus as a Model System at the Intersection of Biology and Physics

Myxococcus xanthus is a unicellular organism whose cells possess the ability to move and communicate, leading to the emergence of complex collective properties and behaviours. This has made it an ideal model system to study the emergence of collective behaviours in interdisciplinary research efforts lying at the intersection of biology and physics, especially in the growing field of active matter research. Often, challenges arise when setting up reliable and reproducible culturing protocols. This tutorial provides a clear and comprehensive guide on the culture, growth, development, and experimental sample preparation of \textit{M. xanthus}. Additionally, it includes some representative examples of experiments that can be conducted using these samples, namely motility assays, fruiting body formation, predation, and elasticotaxis.

physics.bio-ph↗

arXiv2024

Gemini 1.5: Unlocking multimodal understanding across millions of tokens of context

In this report, we introduce the Gemini 1.5 family of models, representing the next generation of highly compute-efficient multimodal models capable of recalling and reasoning over fine-grained information from millions of tokens of context, including multiple long documents and hours of video and audio. The family includes two new models: (1) an updated Gemini 1.5 Pro, which exceeds the February version on the great majority of capabilities and benchmarks; (2) Gemini 1.5 Flash, a more lightweight variant designed for efficiency with minimal regression in quality. Gemini 1.5 models achieve near-perfect recall on long-context retrieval tasks across modalities, improve the state-of-the-art in long-document QA, long-video QA and long-context ASR, and match or surpass Gemini 1.0 Ultra's state-of-the-art performance across a broad set of benchmarks. Studying the limits of Gemini 1.5's long-context ability, we find continued improvement in next-token prediction and near-perfect retrieval (>99%) up to at least 10M tokens, a generational leap over existing models such as Claude 3.0 (200k) and GPT-4 Turbo (128k). Finally, we highlight real-world use cases, such as Gemini 1.5 collaborating with professionals on completing their tasks achieving 26 to 75% time savings across 10 different job categories, as well as surprising new capabilities of large language models at the frontier; when given a grammar manual for Kalamang, a language with fewer than 200 speakers worldwide, the model learns to translate English to Kalamang at a similar level to a person who learned from the same content.

cs.CL↗

arXiv2023

Towards In-context Scene Understanding

In-context learning$\unicode{x2013}$the ability to configure a model's behavior with different prompts$\unicode{x2013}$has revolutionized the field of natural language processing, alleviating the need for task-specific models and paving the way for generalist models capable of assisting with any query. Computer vision, in contrast, has largely stayed in the former regime: specialized decoders and finetuning protocols are generally required to perform dense tasks such as semantic segmentation and depth estimation. In this work we explore a simple mechanism for in-context learning of such scene understanding tasks: nearest neighbor retrieval from a prompt of annotated features. We propose a new pretraining protocol$\unicode{x2013}$leveraging attention within and across images$\unicode{x2013}$which yields representations particularly useful in this regime. The resulting Hummingbird model, suitably prompted, performs various scene understanding tasks without modification while approaching the performance of specialists that have been finetuned for each task. Moreover, Hummingbird can be configured to perform new tasks much more efficiently than finetuned models, raising the possibility of scene understanding in the interactive assistant regime.

cs.CV↗

arXiv2022

Graph Neural Networks for Image Classification and Reinforcement Learning using Graph representations

In this paper, we will evaluate the performance of graph neural networks in two distinct domains: computer vision and reinforcement learning. In the computer vision section, we seek to learn whether a novel non-redundant representation for images as graphs can improve performance over trivial pixel to node mapping on a graph-level prediction graph, specifically image classification. For the reinforcement learning section, we seek to learn if explicitly modeling solving a Rubik's cube as a graph problem can improve performance over a standard model-free technique with no inductive bias.

cs.LG↗

arXiv2018

Equalizer 2.0 - Convergence of a Parallel Rendering Framework

Developing complex, real world graphics applications which leverage multiple GPUs and computers for interactive 3D rendering tasks is a complex task. It requires expertise in distributed systems and parallel rendering in addition to the application domain itself. We present a mature parallel rendering framework which provides a large set of features, algorithms and system integration for a wide range of real-world research and industry applications. Using the Equalizer parallel rendering framework, we show how a wide set of generic algorithms can be integrated in the framework to help application scalability and development in many different domains, highlighting how concrete applications benefit from the diverse aspects and use cases of Equalizer. We present novel parallel rendering algorithms, powerful abstractions for large visualization setups and virtual reality, as well as new experimental results for parallel rendering and data distribution.

cs.GR↗

arXiv2012

Network Bargaining: Using Approximate Blocking Sets to Stabilize Unstable Instances

We study a network extension to the Nash bargaining game, as introduced by Kleinberg and Tardos (STOC'08), where the set of players corresponds to vertices in a graph $G=(V,E)$ and each edge $ij\in E$ represents a possible deal between players $i$ and $j$. We reformulate the problem as a cooperative game and study the following question: Given a game with an empty core (i.e. an unstable game) is it possible, through minimal changes in the underlying network, to stabilize the game? We show that by removing edges in the network that belong to a blocking set we can find a stable solution in polynomial time. This motivates the problem of finding small blocking sets. While it has been previously shown that finding the smallest blocking set is NP-hard (Biro,Kern,Paulusma, TAMC'10), we show that it is possible to efficiently find approximate blocking sets in sparse graphs.

cs.GT↗