Physics-informed neural networks (PINNs) have become a promising framework for simulating partial differential equations (PDEs) by embedding physical laws directly into neural network training. However, recent studies show that PINN optimisation is sensitive to numerical precision. Existing implementations commonly use either single precision (FP32), which is computationally efficient but prone to failure modes, or double precision (FP64), which is robust but substantially expensive. This creates a trade-off between computational efficiency and numerical accuracy. To reduce the computational cost of double-precision training while retaining prediction accuracy, we propose a curvature-aware precision controller that adapts numerical precision during training rather than treating it as a fixed implementation choice. The proposed method reuses curvature information derived from the limited-memory BFGS (L-BFGS) optimiser to construct a precision controller, retaining FP32 when lower precision is sufficient and promoting computation to FP64 when the training dynamics indicate numerical sensitivity or precision-limited stagnation. We evaluate the proposed approach on four one-dimensional PINN failure-mode benchmarks, coupled two-dimensional Navier–Stokes system and Timoshenko beam system, a high-dimensional heat equation, and an irradiance-driven ordinary differential equation example. We further test the proposed approach across different neural network architectures and an additional optimiser. Across the evaluated problems, the method generally achieves accuracy comparable to full FP64 training while reducing training time and providing more reliable convergence than fixed FP32. On the four failure-mode benchmarks, it reaches the mean final FP64 rRMSE target with time-to-target speedups ranging from 1.10× to 2.13×. The obtained results suggest that precision sensitivity in PINN optimisation is phase-dependent and that selectively applying higher precision during numerically critical stages can lower computational cost while preserving accuracy close to full FP64 training.

Farmers rely on in-field observations to make well-informed crop management decisions to maximize profit and minimize adverse environmental impact. However, obtaining real-world crop state measurements is labor-intensive, time-consuming and expensive. In most cases, it is not feasible to gather crop state measurements before every decision moment. Moreover, in previous research pertaining to farm management optimization, these observations are often assumed to be readily available without any cost, which is unrealistic. Hence, enabling optimization without the need to have *temporally complete* crop state observations is important. An approach to that problem is to include measuring as part of decision making. As a solution, we apply reinforcement learning (RL) to recommend opportune moments to simultaneously measure crop features and apply nitrogen fertilizer. With realistic considerations, we design an RL environment with explicit crop feature measuring costs. While balancing costs, we find that an RL agent, trained with recurrent PPO, discovers adaptive measuring policies that follow critical crop development stages, with results aligned by what domain experts would consider a sensible approach. Our results highlight the importance of measuring when crop feature measurements are not readily available.

Biophysical models offer valuable insights into climate-phenology relationships in both natural and agricultural settings. However, there are substantial structural discrepancies across models which require site-specific recalibration, often yielding inconsistent predictions under similar climate scenarios. Machine learning methods offer data-driven solutions, but often lack interpretability and alignment with existing knowledge. We present a phenology model describing dormancy in fruit trees, integrating conventional biophysical models with a neural network to address their structural disparities. We evaluate our hybrid model in an extensive case study predicting cherry tree phenology in Japan, South Korea and Switzerland. Our approach consistently outperforms both traditional biophysical and machine learning models in predicting blooming dates across years. Additionally, the neural network's adaptability facilitates parameter learning for specific tree varieties, enabling robust generalization to new sites without site-specific recalibration. This hybrid model leverages both biophysical constraints and data-driven flexibility, offering a promising avenue for accurate and interpretable phenology modeling.

In contrast to the rapid digitalization of several industries, agriculture suffers from low adoption of smart farming tools. Even though recent advancements in AI-driven digital agriculture can offer high-performing predictive functionalities, they lack tangible quantitative evidence on their benefits to the farmers. Field experiments can derive such evidence, but are often costly, time consuming and hence limited in scope and scale of application. To this end, we propose an observational causal inference framework for the empirical evaluation of the impact of digital tools on target farm performance indicators (e.g., yield in this case). This way, we can increase farmers' trust via enhancing the transparency of the digital agriculture market, and in turn accelerate the adoption of technologies that aim to secure farmer income resilience and global agricultural sustainability against a changing climate. As a case study, we designed and implemented a recommendation system for the optimal sowing time of cotton based on numerical weather predictions, which was used by a farmers' cooperative during the growing season of 2021. We then leverage agricultural knowledge, collected yield data, and environmental information to develop a causal graph of the farm system. Using the back-door criterion, we identify the impact of sowing recommendations on the yield and subsequently estimate it using linear regression, matching, inverse propensity score weighting and meta-learners. The results revealed that a field sown according to our recommendations exhibited a statistically significant yield increase that ranged from 12% to 17%, depending on the method. The effect estimates were robust, as indicated by the agreement among the estimation methods and four successful refutation tests. We argue that this approach can be implemented for decision support systems of other fields, extending their evaluation beyond a performance assessment of internal functionalities.

SUMMARY: ChromaX is a Python library that enables the simulation of genetic recombination, genomic estimated breeding value calculations, and selection processes. By utilizing GPU processing, it can perform these simulations up to two orders of magnitude faster than existing tools with standard hardware. This offers breeders and scientists new opportunities to simulate genetic gain and optimize breeding schemes. AVAILABILITY AND IMPLEMENTATION: The documentation is available at https://chromax.readthedocs.io. The code is available at https://github.com/kora-labs/chromax.

To improve the sustainability and resilience of modern food systems, designing improved crop management strategies is crucial. The increasing abundance of data on agricultural systems suggests that future strategies could benefit from adapting to environmental conditions, but how to design these adaptive policies poses a new frontier. A natural technique for learning policies in these kinds of sequential decision-making problems is reinforcement learning (RL). To obtain the large number of samples required to learn effective RL policies, existing work has used mechanistic crop growth models (CGMs) as simulators. These solutions focus on single-year, single-crop simulations for learning strategies for a single agricultural management practice. However, to learn sustainable long-term policies we must be able to train in multi-year environments, with multiple crops, and consider a wider array of management techniques. We introduce CYCLESGYM, an RL environment based on the multi-year, multi-crop CGM Cycles. CYCLESGYM allows for long-term planning in agroecosystems, provides modular state space and reward constructors and weather generators, and allows for complex actions. For RL researchers, this is a novel benchmark to investigate issues arising in real-world applications. For agronomists, we demonstrate the potential of RL as a powerful optimization tool for agricultural systems management in multi-year case studies on nitrogen (N) fertilization and crop planning scenarios.

Big Data has been seen as a remedy for the efficient management of the ever-increasing genomic data. In this paper, we investigate the use of Apache Spark to store and process Variant Calling Files (VCF) on a Hadoop cluster. We demonstrate Tomatula, a software tool for converting VCF files to Apache Parquet storage format, and an application to query variant calling datasets. We evaluate how the wall time (i.e. time until the query answer is returned to the user) scales out on a Hadoop cluster storing VCF files, either in the original flat-file format, or using the Apache Parquet columnar storage format. Apache Parquet can compress the VCF data by around a factor of 10, and supports easier querying of VCF files as it exposes the field structure. We discuss advantages and disadvantages in terms of storage capacity and querying performance with both flat VCF files and Apache Parquet using an open plant breeding dataset. We conclude that Apache Parquet offers benefits for reducing storage size and wall time, and scales out with larger datasets.

We present Pythia, a privacy-enhanced non-invasive contextual suggestion system for tourists, with important architectural innovations. The system offers high quality personalized recommendations, non-invasive operation and protection of user privacy. A key feature of Pythia is the exploitation of the vast amounts of personal data generated by smartphones to automatically build user profiles, and make contextual suggestions to tourists. More precisely, the system utilizes (sensitive) personal data, such as location traces, browsing history and web searches (query logs), to build a POI-based user profile. This profile is then used by a contextual suggestion engine for making POI recommendations to the user based on her current location. Strong privacy guarantees are achieved by placing both mechanisms at the user-side. As a proof of concept, we present a Pythia prototype which implements the aforementioned mechanisms as mobile applications for Android, as well as, web applications.

The limited reuse of current environmental software can be blamed in part on the tools used to develop it; the use of generic-purpose programming languages makes it particularly hard. As environmental scientists strive to prioritize the clear statement and communication of the semantics of natural systems in favor of understanding software implementations of their models, Domain-Specific Languages may come to help, offering the option of truly declarative environment for environmental modeling. This paper discusses some key requirements and concepts for developing Domain-Specific Languages that can inform and streamline environmental modeling, and previews some use scenarios using examples from a DSL in development.

Participatory sensing is a crowd-sourcing technique which relies both on active contribution of citizens and on their location and mobility patterns. As such, it is particularly vulnerable to privacy concerns, which may seriously hamper the large-scale adoption of participatory sensing applications. In this paper, we present a privacy-preserving system architecture for participatory sensing contexts which relies on cryptographic techniques and distributed computations in the cloud. Each individual is represented by a personal software agent, which is deployed on one of the popular commercial cloud computing services. The system enables individuals to aggregate and analyse sensor data by performing a collaborative distributed computation among multiple agents. No personal data is disclosed to anyone, including the cloud service providers. The distributed computation proceeds by having agents execute a cryptographic protocol based on a homomorphic encryption scheme in order to aggregate data. We show formally that our architecture is secure in the Honest-But-Curious model both for the users and the cloud providers. Our approach was implemented and validated on top of the NoiseTube system [1], [2], which enables participatory sensing of noise. In particular, we repeated several mapping experiments carried out with NoiseTube, and show that our system is able to produce identical outcomes in a privacy-preserving way. We experimented with real and simulated data, and present a live demo running on a heterogeneous set of commercial cloud providers. The results show that our approach goes beyond a proof-of-concept and can actually be deployed in a real-world setting. To the best of our knowledge this system is the first operational privacy-preserving approach for participatory sensing. While validated in terms of NoiseTube, our approach is useful in any setting where data aggregation can be performed with efficient homomorphic cryptosystems.

Knowledge-based software engineering enables a programmer to integrate rich semantics in the software development process. In this work, we show how an OWL/RDF knowledge base can be integrated with conventional domain-centric data models (enterprise Java beans) and object-relational mapping toolkits (Hibernate). We present a pathway for the software developer to generate enterprise Java beans source code and hibernate object-relational mappings starting from a domain ontology. This way, a semantic-rich enterprise development environment is specified that combines the benefits of using ontologies with software development standards.

Accurate and on-line decision-making is required by decision support systems including those ones used for environmental information management. This paper focuses on air quality assessment and demonstrates the added value of applying data mining techniques in operational decision-making. More specifically, the application of fuzzy lattice reasoning (FLR) classifier is investigated. An enhanced FLR learning algorithm is presented that employs a sigmoid valuation function for introducing tunable non-linearities. The FLR classifier is applied here beyond the unit-hypercube. The FLR with a sigmoid positive valuation function demonstrates an improved performance on a dataset from the region of Valencia, Spain regarding an environmental problem. Descriptive decision making knowledge (i.e. rules) for classification is also induced.