Diabetic retinopathy grading is inherently ordinal and long-tailed, with minority stages being scarce, heterogeneous, and clinically critical to detect accurately. Conventional methods often rely on isotropic Gaussian priors and symmetric loss functions, misaligning latent representations with the task's asymmetric nature. We propose the Constrained Asymmetric Prior Wasserstein Autoencoder (CAP-WAE), a novel framework that addresses these challenges through three key innovations. Our approach employs a Wasserstein Autoencoder (WAE) that aligns its aggregate posterior with a asymmetric prior, preserving the heavy-tailed and skewed structure of minority classes. The latent space is further structured by a Margin-Aware Orthogonality and Compactness (MAOC) loss to ensure grade-ordered separability. At the supervision level, we introduce a direction-aware ordinal loss, where a lightweight head predicts asymmetric dispersions to generate soft labels that reflect clinical priorities by penalizing under-grading more severely. Stabilized by an adaptive multi-task weighting scheme, our end-to-end model requires minimal tuning. Across public DR benchmarks, CAP-WAE consistently achieves state-of-the-art Quadratic Weighted Kappa, accuracy, and macro-F1, surpassing both ordinal classification and latent generative baselines. t-SNE visualizations further reveal that our method reshapes the latent manifold into compact, grade-ordered clusters with reduced overlap.

Applying AI foundation models directly to geospatial datasets remains challenging due to their limited ability to represent and reason with geographical entities, specifically vector-based geometries and natural language descriptions of complex spatial relations. To address these issues, we investigate the extent to which a well-known-text (WKT) representation of geometries and their spatial relations (e.g., topological predicates) are preserved during spatial reasoning when the geospatial vector data are passed to large language models (LLMs) including GPT-3.5-turbo, GPT-4, and DeepSeek-R1-14B. Our workflow employs three distinct approaches to complete the spatial reasoning tasks for comparison, i.e., geometry embedding-based, prompt engineering-based, and everyday language-based evaluation. Our experiment results demonstrate that both the embedding-based and prompt engineering-based approaches to geospatial question-answering tasks with GPT models can achieve an accuracy of over 0.6 on average for the identification of topological spatial relations between two geometries. Among the evaluated models, GPT-4 with few-shot prompting achieved the highest performance with over 0.66 accuracy on topological spatial relation inference. Additionally, GPT-based reasoner is capable of properly comprehending inverse topological spatial relations and including an LLM-generated geometry can enhance the effectiveness for geographic entity retrieval. GPT-4 also exhibits the ability to translate certain vernacular descriptions about places into formal topological relations, and adding the geometry-type or place-type context in prompts may improve inference accuracy, but it varies by instance. The performance of these spatial reasoning tasks offers valuable insights for the refinement of LLMs with geographical knowledge towards the development of geo-foundation models capable of geospatial reasoning.

Drug discovery remains a slow and expensive process that involves many steps, from detecting the target structure to obtaining approval from the Food and Drug Administration (FDA), and is often riddled with safety concerns. Accurate prediction of how drugs interact with their targets and the development of new drugs by using better methods and technologies have immense potential to speed up this process, ultimately leading to faster delivery of life-saving medications. Traditional methods used for drug-target interaction prediction show limitations, particularly in capturing complex relationships between drugs and their targets. As an outcome, deep learning models have been presented to overcome the challenges of interaction prediction through their precise and efficient end results. By outlining promising research avenues and models, each with a different solution but similar to the problem, this paper aims to give researchers a better idea of methods for even more accurate and efficient prediction of drug-target interaction, ultimately accelerating the development of more effective drugs. A total of 180 prediction methods for drug-target interactions were analyzed throughout the period spanning 2016 to 2025 using different frameworks based on machine learning, mainly deep learning and graph neural networks. Additionally, this paper discusses the novelty, architecture, and input representation of these models.

Company executives will host a roundtable discussion, present to session attendees and participate on a panel about how neuromorphic computing addresses the challenges at the edge. The Edge AI Summit aids enterprise adopters, OEMs, AI software and hardware providers in their pursuit of deploying AI at the edge by helping them form partnerships that will improve the performance of embedded hardware, optimize models and compress workloads to enable high-performing and low-footprint AI to exist at the edge. "With the acceleration of AI at the edge, what better place to demonstrate the capabilities of our Akida advanced neuromorphic architecture than at the Edge AI Summit," said Mankar. "We are at the stage where evangelizing our approach to this emerging market opportunity is being replaced by our introduction of a production version of the Akida chip. I know Rob and I are eager to present and discuss the possibilities of AI at the edge as well as how Akida will revolutionize the industry in a way that existing technologies are not capable of."

WIRE)--BrainChip Holdings Ltd (ASX: BRN), (OTCQX: BRCHF), a leading provider of ultra-low power high performance artificial intelligence technology, today announced BrainChip will be taking orders of two development kits for its Akida advanced neural networking processor, enabling partners, large enterprises, and OEMs to begin internal testing and validation of Akida's high-performance, small, ultra-low power AI chip. Akida NSoC and intellectual property enable a wide array of edge AI capabilities that include continuous learning and inference. BrainChip is offering two development kits, both including the AKD1000 chip on a mini-PCI board: an X86 Shuttle PC development kit, as well as an ARM-based Raspberry Pi development kit. "Offering development kits is not only a major step towards full commercialization, it's also an exciting opportunity to see how our partners and future customers will put Akida to work in environments and scenarios like consumer electronics, industrial applications, aerospace and defense systems, healthcare and medical devices, automotive technology, and more," said Anil Mankar, BrainChip co-founder and chief development officer. "We believe the AKD1000 silicon, or the licensing of Akida in a configurable IP format, will lead to major changes in industries using AI at the edge because of its performance, security, low power requirements, and mainly Akida's ability to perform AI training and learning on the device itself, without dependency on the cloud."

Akida NSoC and intellectual property enable a wide array of edge AI capabilities that include continuous learning and inference. BrainChip is offering two development kits both including the AKD1000 chip on a mini-PCI board: an X86 Shuttle PC development kit as well as an ARM-based Raspberry Pi development kit. "Offering development kits is not only a major step towards full commercialization, it's also an exciting opportunity to see how our partners and future customers will put Akida to work in environments and scenarios like consumer electronics, industrial applications, aerospace and defense systems, healthcare and medical devices, automotive technology, and more," said Anil Mankar, BrainChip co-founder and chief development officer. "We believe the AKD1000 silicon, or the licensing of Akida in a configurable IP format, will lead to major changes in industries using AI at the edge because of its performance, security, low power requirements, and mainly Akida's ability to perform AI training and learning on the device itself, without dependency on the cloud." Development kits for Akida-based applications and solutions evolving to production status are a step toward joining the neuromorphic revolution for edge AI applications.

ALISO VIEJO, Calif., August 26, 2021--(BUSINESS WIRE)--BrainChip Holdings Ltd (ASX: BRN), (OTCQX: BRCHF), a leading provider of ultra-low power high performance artificial intelligence technology, was recognized as one of the "Startups Worth Watching in 2021" in EE Times' annual Silicon 100 list of global semiconductor technologies. EE Times' 21st revision of the Silicon 100 tracks the pulse of the industry to identify emerging technology trends and developments that hold promise for the future. This year, the publication chose to analyze the Silicon 100 in more detail with 22 categories that run from materials and packaging at a fundamental extreme to quantum computing and security at the highest level of abstraction. BrainChip was recognized in the "Specialist (GPU-Through-AI) Processor, Accelerators" category. Selection of companies to the Silicon 100 is based on criteria including technology, intended market, financial position and investment profile, maturity and executive leadership.

The chips were manufactured at Taiwan Semiconductor Manufacturing Company (TSMC) from a production mask set provided in May 2021. This mask set follows the successful production of engineering samples from the Company's Multi-Project Wafers (MPW), received in August of 2020, and the subsequent delivery of evaluation boards. SNA supported all assembly and test operations for the Akida devices, including a review of the TSMC Process Control Monitoring (PCM) data, assembly, device electrical testing, and simulation correlation. These Akida devices will support BrainChip's Early Access Program (EAP) customers, future customers with whom the Company has engaged using the BrainChip software development environment MetaTF and others which have existing Convolutional Neural Networks (CNNs) and seek performance improvements in terms of power consumption, design flexibility, and true learning at the edge. "As a company our goal has been to put our chips and IP into the hands of customers and partners so they in turn can transform edge AI for implementations like home automation, industrial IoT, security and cybersecurity, autonomous vehicles, medical devices, and leveraging sensor technology for objects, sound, odor, taste, vibration and more," said Rob Telson, BrainChip vice president of sales and marketing.

WIRE)--BrainChip Holdings Ltd (ASX: BRN), (OTCQX: BRCHF), a leading provider of ultra-low power high performance artificial intelligence technology, today announced it has received the first batch of Akida chips from its manufacturing run from Socionext America (SNA). The chips were manufactured at Taiwan Semiconductor Manufacturing Company (TSMC) from a production mask set provided in May 2021. This mask set follows the successful production of engineering samples from the Company's Multi-Project Wafers (MPW), received in August of 2020, and the subsequent delivery of evaluation boards. SNA supported all assembly and test operations for the Akida devices, including a review of the TSMC Process Control Monitoring (PCM) data, assembly, device electrical testing, and simulation correlation. These Akida devices will support BrainChip's Early Access Program (EAP) customers, future customers with whom the Company has engaged using the BrainChip software development environment MetaTF and others which have existing Convolutional Neural Networks (CNNs) and seek performance improvements in terms of power consumption, design flexibility, and true learning at the edge.

ALISO VIEJO, Calif., June 28, 2021--(BUSINESS WIRE)--BrainChip Holdings Ltd (ASX: BRN), (OTCQX: BRCHF), a leading provider of ultra-low power, high-performance artificial intelligence technology, today announced the appointment of Antonio J. Viana as a non-executive director. He joins company founder and CEO Peter van der Made and fellow non-executive directors Emmanuel T. Hernandez and Geoffry Carrick on BrainChip's Board of Directors. "Antonio is a very welcome addition to our Board and brings with him outstanding industry knowledge, contacts and experience," said CEO Peter van der Made. "He will play a substantial and influential role on our Board as BrainChip transitions from an R&D focus to sales and production of our Akida neuromorphic artificial intelligence technology." Mr. Viana currently serves as the executive chairman at QuantalRF AG, an emerging next-generation, front-end RF company developing transformative wireless communication solutions.