Sachin Abeywardana, PhD

Excited to announce that Studitory is partnering with CSIRO's Data61, working closely with Xiwei (Sherry) Xu and the Applied AI systems team to develop a cutting-edge AI marking workflow to mimic real teacher marking patterns! This collaboration will enable us to fully automate the exam practice experience on Studitory, making learning smarter, more reliable and more personalized than ever. A huge thank you to Sherry and the Data61 team for joining us on this journey, excited to see how we can transform the learning space! #AI #EdTech #Innovation #Studitory #Data61 #Collaboration

30

1 Comment

Proud to have contributed to our collaborative project with Google Cloud Security on open-source software security! We've developed key tools like PROBER and a Vulnerability Prioritisation approach to strengthen the software supply chain. #SoftwareSecurity #SupplyChainSecurity #VulnerabilityManagement #CSIRO #Google

10

1 Comment

Not all model outputs are created equal. Exploring a great tool for quantifying uncertainty in LLMs. As we integrate Large Language Models (LLMs) into more critical applications, understanding when to trust their output is just as important as the output itself. This is where Uncertainty Quantification (UQ) becomes essential. Open-source library called uQLM (Uncertainty Quantification for Language Models) that tackles this exact problem. It provides a comprehensive toolkit to measure an LLM's confidence in its own generations, going beyond a simple softmax probability. The repo includes implementations for several state-of-the-art methods: > Semantic entropy for detecting uncertainty in free-form generations > Pragmatic entropy and Pragmatic calibration > Ensemble-based methods This is crucial for building reliable AI systems in high-stakes fields like healthcare, finance, or legal, where a model's miscalibrated confidence can lead to significant errors. By quantifying doubt, we can build safeguards, trigger human review, and ultimately create more trustworthy and responsible AI. Major kudos to the AI team at CVS Health for open-sourcing this valuable contribution to the ML community. It's a concrete step towards more robust and transparent AI. Check out the repository here: https://lnkd.in/gHNwXEjb #AI #MachineLearning #LLM #UncertaintyQuantification #ResponsibleAI #ExplainableAI #XAI #MLOps #ArtificialIntelligence #OpenSource #HealthTech

10

Well, I’ve been complaining that Australia has no AI Strategy but then I’d be pretty disappointed if it turned out like NZ’s strategy. Supporting AI adoption without building up your capability to create your own AI is putting your future in the hands of other nations. #sovereignai #techmadeinaustralia Chamara Somaratne Eddie Major

25

7 Comments

New blog post: Using deep learning in PyTorch on rainfall data to predict Australian area remoteness Who would have thought I'd be back with another post so soon! Not me, that's for sure. Anyway, in this episode, I continue my exploration of rainfall data for each ABS SA2 in Australia by computing a set of time-series features and then using them to train deep neural networks in PyTorch to classify whether an unseen rainfall time series is from a major city, inner/outer regional Australia, or remote/very remote Australia. What's more, I switch between R and Python within the one document, highlighting the power of multilingual integration in modern markdown! I found that time-series features which capture temporal dynamics in rainfall can meaningfully classify the remoteness of ABS SA2s, which suggests that there are differences in temporal patterns of rainfall unique to the five different remoteness classifications. This is interesting, because it highlights that looking beyond simple/intuitive hypotheses such as 'cities probably get more rain than the outback' can reveal potentially novel insights about the differential patterns. This could lead to further inquiry about whether these dynamics shift over time (i.e., due to climate change) or other important scientific questions. In any case, if nothing else, deep learning is awesome and I had fun! Please check out the post via the link below if you are interested. https://lnkd.in/giQ-UHKR #statistics #deeplearning #machinelearning #climate #timeseries

8

📄 Paper Alert! Excited to share another recent publication, “AI for All: Identifying AI Incidents Related to Diversity and Inclusion” by CSIRO's Data61 AI Diversity and Inclusion team, Rifat Ara Shams, PhD, Prof Didar Zowghi, and Muneera Bano, in the Journal of Artificial Intelligence Research. Link to the paper🔗 https://lnkd.in/gSASKnfZ In this work, we highlight why it is critical to systematically identify and study AI incidents through the lens of D&I. Our manual analysis of two major AI incident databases revealed: ✅ 34.3% of 551 incidents from the AI Incident Database (AIID - https://lnkd.in/g7YE4Dfi) were related to D&I issues. ✅ 46.45% of 310 incidents from the AIAAIC database (https://www.aiaaic.org/) were related to D&I issues. These incidents reflected harms connected to attributes such as race, gender, age, socio-economic status, religion, and more. To support the responsible development of AI systems, we proposed a decision tree to guide the identification of D&I-related incidents and created a public repository for further research and action. #AIethics #ResponsibleAI #DiversityAndInclusion #AIIncidents #Fairness

40

Interesting news for open science and genomics! UNSW Sydney, Pawsey, and AMD have introduced Slorado — an open-source, real-time nanopore DNA basecaller. Slorado accelerates whole-genome analysis and expands open, high-performance genomics for researchers worldwide. The Setonix supercomputer at the Pawsey Supercomputing Research Centre, powered by AMD Instinct GPUs, served as the primary testbed for Slorado’s performance validation. See the article for more details! #AMD #Genomics #OpenScience #HPC #ROCm #AMDBrandAmbassador

4

How did a university side project become a model for AI at scale?   At Data & AI Edge on 1 April, Professor Danny Liu from the University of Sydney will share the story behind Cogniti, an AI platform that began as a grassroots solution and now supports over 2,500 agents across Australia.   With a rare blend of technical depth and education-sector insight, Danny will unpack how a simple idea grew into a scalable, human-centred AI framework.   He will explore how empowering domain experts, building a culture of innovation, and designing flexible governance enabled AI to move from pilot to production across sectors. #DataAIEdge

2

1 Comment

🔥𝐇𝐍𝐒𝐖 is the most widely used vector indexing algorithm. Compared to other popular index types such as IVF and SCaNN, it provides better search latency at similar level of recall. 𝐇𝐍𝐒𝐖 𝐢𝐬 𝐬𝐡𝐨𝐫𝐭 𝐟𝐨𝐫 𝐇𝐢𝐞𝐫𝐚𝐫𝐜𝐡𝐢𝐜𝐚𝐥 𝐍𝐚𝐯𝐢𝐠𝐚𝐛𝐥𝐞 𝐒𝐦𝐚𝐥𝐥 𝐖𝐨𝐫𝐥𝐝. Inspired by Navigable Small World (NSW), it is a graph-based index and its innovation is a multi-layered graph structure on NSW that significantly improves search efficiency. In HNSW, vectors are organized into a series of layers. The top layers contain a sparse set of vectors and serve as entry points for the search. As the algorithm descends into lower layers, the density of connections increases, allowing for more accurate local navigation. This hierarchical design enables 𝐟𝐚𝐬𝐭 𝐚𝐩𝐩𝐫𝐨𝐱𝐢𝐦𝐚𝐭𝐞 𝐧𝐞𝐚𝐫𝐞𝐬𝐭 𝐧𝐞𝐢𝐠𝐡𝐛𝐨𝐫 (𝐀𝐍𝐍) 𝐬𝐞𝐚𝐫𝐜𝐡 with logarithmic time complexity. Here are 3 concepts to help you better understand HNSW.👇 Want to learn why HNSW can achieve high recall at relative low latency? Check out https://lnkd.in/gxVx6Fca —————— Follow Milvus, created by Zilliz to learn more. #HNSW #VectorSearch #RAG #Milvus

11

Valyu's CEO, Hirsh Pithadia, has just made an appearance on Jennifer Ding's "AI Data Chats"! In the episode, Hirsh lays out a strong case for why most current approaches to retrieval, especially 'context stuffing', are hitting their limits. He gets into: 💻Why data infrastructure, not just better prompts, drives performant AI 🤖 How we should think about building for the coming wave of agentic AI 💸 The overlooked economics of context delivery at inference time If you’re building anything in the RAG, infra, or agent space, this one’s worth a listen. 🎙️ Watch the full episode here: https://lnkd.in/eWvNzv36

16

1 Comment

Month on month, we’re seeing a sharp rise in clients adopting Generative AI across cloud ecosystems: From early stage proofs of concept to enterprise scale deployments on Azure, AWS, and GCP. With this surge comes a wave of new roles in GenAI engineering, MLOps, and AI infrastructure, and the battle for talent is well and truly on. - For top engineers and architects, this is the moment to be in the market. - For companies, the challenge is clear: move fast, communicate value, and attract the scarce skills driving the next wave of innovation. At Nuage, we’re proud to be partnering with forward thinking exec teams shaping the future of Cloud and AI in Australia.

4

What really stuck with me from this interview with Manos Raptopoulos: “𝐂𝐮𝐬𝐭𝐨𝐦𝐞𝐫𝐬 𝐰𝐚𝐧𝐭 𝐨𝐮𝐭𝐜𝐨𝐦𝐞𝐬, 𝐧𝐨𝐭 𝐞𝐱𝐩𝐞𝐫𝐢𝐦𝐞𝐧𝐭𝐬.” That’s exactly what I’m seeing across Australia and New Zealand. That there’s strong appetite for AI, but only where it’s 𝐠𝐫𝐨𝐮𝐧𝐝𝐞𝐝 𝐢𝐧 𝐫𝐞𝐚𝐥 𝐛𝐮𝐬𝐢𝐧𝐞𝐬𝐬 𝐯𝐚𝐥𝐮𝐞, trusted data and clear accountability. There’s a refreshing pragmatism here: AI adoption isn’t about speed for speed’s sake… it’s about 𝐜𝐥𝐚𝐫𝐢𝐭𝐲, 𝐭𝐫𝐮𝐬𝐭 𝐚𝐧𝐝 𝐨𝐮𝐭𝐜𝐨𝐦𝐞𝐬. What resonates most is the focus on:  • Business-case driven AI  • Enterprise-grade data foundations  • Precision and governance in the “last mile” This is how AI earns trust, and delivers value, at scale. Check it out: 🔗 https://lnkd.in/gApb39hw

11

🚨 Released today, this data-driven update paints a detailed picture of how AI is evolving in Australia... Just released by CSIRO and the National AI Centre — the 2025 report on Australia’s Artificial Intelligence Ecosystem: Growth and Opportunities. Here’s what stood out: ➡️ AI jobs are surging: over 1,500 organisations sought AI-skilled workers in 2024, up from just 483 in 2015 ➡️ Research output doubled and patents quadrupled since 2015. But, there’s a lag in commercialisation with a 23:1 ratio of research publications to patents ➡️ Strong growth in small enterprise-led innovation, but the ecosystem remains fragile. 85% of AI companies employ fewer than 50 staff ➡️ Urban AI clusters are thriving (Melbourne, Sydney, Brisbane, Perth). Each with specialised capabilities from healthcare to mining. One of the highlights follows on from my previous posts about AI sovereignty. The report sees Australia’s dual identity as both AI taker and AI maker as more than a label — it’s a strategic stance: 🔧 AI taker: rapidly adopting global tools to optimise operations, deliver smarter services, and lift productivity 🌱 AI maker: building tailored AI solutions in high-potential sectors from precision agriculture to remote healthcare and mining automation 🎯 Strength lies in contextual application adapting AI to fit our unique societal, environmental, and industry conditions 🧬 Innovation corridors are emerging with research and industry co-locating to drive region-specific AI capability ❓It's a particularly positive spin on Australia's current sovereign AI position. But I think the big question remains, how do we turn this hybrid position into long-term impact and sovereign capability? Read the full report 👉 https://lnkd.in/gKKFBcYP #ArtificialIntelligence #AIEcosystem #TechPolicy #Innovation #ResponsibleAI #NAIC #CSIRO #AustraliaAI Disclaimer: GenAI helped me generate this content.

49

3 Comments

Last month, CHeBA’s health data scientist Rory Chen delivered an engaging workshop with R-Ladies Global on using #Shiny to build interactive apps and dashboards. The session highlighted the potential of interactive data visualisation to make complex research accessible and impactful - transforming datasets into user-friendly insights that can be explored by clicking, scrolling, and swiping. Rory is also the lead creator of the Data Repository Explorer (DataRepExp), an open-source Shiny app published in the Journal of Open Source Software. Designed to simplify access to research data, it enables high-level comparison across multiple studies - addressing major challenges in data access and harmonisation. Workshops like Rory’s demonstrate what’s possible when researchers are equipped with the right tools. With targeted infrastructure, we can scale this capacity and strengthen the bridge between research and real-world impact. 🔗 Watch now: https://lnkd.in/gV8GiAVa

6

Surprising that the report was "AI generated" without human-in-the-loop verification: https://lnkd.in/geFyMD2p Companies should wake up to the fact that AI will hallucinate - it is a Godelian inevitability. This is primarily due to the training process, where they are trained to answer questions, without verification of whether those answers are correct or not. This is similar to a student who guesses answers in a test - this is fine as long as there are no negative marks, which can cause the student to get significantly lower marks than otherwise. This is similar to AI hallucinations, which are fundamentally inevitable, as this study shows (link in the comment below).

16

1 Comment

☢️ Ask your AI to “pick a number between 1 and 50.” 🚨 Here’s what we just found out: Ask your AI to “pick a number between 1 and 50.” Did it say 27? All major models (#OpenAI, #Claude, etc.) do too. This isn’t true #randomness: It’s a learned #statistical_echo. #Why? #AI selects #outputs by computing token probabilities (logits). "#27" has a high learned #probability from human text patterns. The output isn’t a #creative decision, it’s the most likely next token. If you doubt it, check your logit distributions. The spike around 27 will prove it. ✅ #LoopHole addressed. Just another #reminder: most AI “#choices” are well-polished #averages, not #unique thoughts. 👉 What does this mean for #work? As we #adopt AI for critical #decisions and #creative #tasks, remember: these systems mirror human #patterns rather than #think independently. #Blind trust in AI outputs can reinforce #biases and #limit true #innovation. Always #audit, #question and #guide - don't just #automate. Contact for your AI Applied Solutions and Concerns or Training: https://lnkd.in/gREkEGgj This is how #average and normal the present #AI models are: And that's a dire concern for every Agentic AI #investment and #deployments. #Meta , #Claude #Grok etc)

1

🎙️ Ryan Donovan and Benjamin Popper welcome Douwe Kiela, CEO and cofounder of Contextual AI, to the podcast to explore RAG, Douwe's early research at Meta that jumpstarted it, and the challenges of hallucinations. https://lnkd.in/gb9BeAfn

52

1 Comment

Delighted to share that the Australian Government is establishing a new AI Safety Institute (AISI). The AISI will provide trusted, expert capability to monitor, test and share information on emerging AI technologies, risks and harms, and will become operational in 2026. There is a huge amount of work ahead to build this capability, and I am excited to work with colleagues and partners across government, industry and academia to help make it a reality. By identifying future risks, the AISI will help ensure fit-for-purpose protections for Australians while supporting the safe and responsible development of AI across our economy and society. #AISafety https://lnkd.in/gTUSyQgH

120

13 Comments