Tharindu Kumarage

📣 Call for Papers: #TrustNLP2026 is now open! We're seeking submissions on trustworthy, ethical, explainable & robust NLP. 📅 Paper deadline: March 5, 2026 Submit & details ➡️ trustnlpworkshop.github.io #NLP #AI #ResponsibleAI

We've been hard at work solving a persistent industry problem: frontier models launch with impressive benchmarks, organizations test them, then they don't work for actual needs. To help bridge this gap, we're excited to announce Amazon Nova Forge – a new way to build frontier AI models that are experts in your domain.

Amazon is back with Nova 2.0, a substantial upgrade over prior Amazon Nova models and demonstrating particular strength in agentic capabilities Amazon has released Nova 2.0 Pro (Preview), its new flagship model; Nova 2.0 Lite, focused on speed and lower cost; and Nova 2.0 Omni, a multimodal model handling text, image, video and speech inputs with text and image outputs. Key benchmarking takeaways: Amazon back amongst top AI players: This is Amazon’s latest release since Nova Premier and Amazon’s first release of reasoning models. Nova 2.0 Pro jumps 30 points in the Artificial Analysis Intelligence Index over Premier and Lite 38 points. This represents a huge increase in capabilities and Amazon’s return to being amongst the top AI players. Strengths in agentic capabilities: Agentic capabilities including tool calling is a strength of the models, Nova 2.0 Pro scores 93% on τ²-Bench Telecom and 80% on IFBench on medium and high reasoning budgets respectively (complete benchmarks for high reasoning coming soon). This places Nova 2.0 Pro Preview amongst the leading models in these benchmarks. Multimodal: Nova 2.0 Omni is one of few models, alongside most notably the Gemini model series, that can natively handle text, image, video and speech inputs. This is a new differentiator for Amazon’s Nova model series. Competitive pricing: Amazon has priced Nova 2.0 Pro at $1.25/$10 per million input/output tokens, and considering token usage the model took $662 to run our Artificial Analysis Intelligence Index. This is substantially less than other frontier models like Claude 4.5 Sonnet ($817) and Gemini 3 Pro ($1201), but remains above others including Kimi K2 Thinking ($380). Nova 2.0 Lite and Omni are both priced at $0.3/$2.5 per million input/output tokens. See below for further analysis

We're thrilled to announce that the Fifth Workshop on Trustworthy NLP (TrustNLP) is coming to #NAACL this year! 🥳 🚀 The Call for Papers (CFP) is now live—don’t miss your chance to contribute! Stay tuned for updates and visit our website: trustnlpworkshop.github.io

🚨 Thrilled to announce our EMNLP 2024 paper on defending against LLM-driven social engineering attacks, co-authored with @TharinduKTS, @Amrita_Bh, and more! Check it out: arxiv.org/pdf/2406.12263 🚀 #EMNLP2024 #NLP

Mindful-RAG: A Study of Points of Failure in Retrieval Augmented Generation Identifies key issues in knowledge graph-based RAG systems for LLMs, focusing on question intent and context alignment. 📝arxiv.org/abs/2407.12216

We are excited to share that our paper, "Can Knowledge Graphs Reduce Hallucinations in LLMs? : A Survey" is accepted in main conference at NAACL 2024. Do checkout the latest paper @ arxiv.org/abs/2311.07914 @TharinduKTS @liuhuan #NAACL2024

📝 Announcing our new survey paper on "AI-generated Text Forensic Systems: Detection, Attribution, and Characterization" 🔹 In collaboration with @ASU 🔹 PDF: arxiv.org/pdf/2403.01152… #ArtificialInteligence #AI #MachineLearning #LLMs

Gihan Ketawala et al.: The Pixel Anomaly Detection Tool: a user-friendly GUI for classifying detector frames using machine-learning approaches #MachineLearning #SerialCrystallography #ImageClassification @asu... #IUCr scripts.iucr.org/cgi-bin/paper?…

@Anuradhawick @PasinduTennage IMO, this is what adopting a new technology looks like. Educational institutes are also looking into leveraging AI for teaching and learning activities rather than fighting against misuse cases. Similarly, the interview process will evolve as well.

@PasinduTennage Remember back then some interviews tested us on hackerrank? Algorithms etc. I think now questions should probably level up and ask to solve complete problems. Not small algorithm implementations or optimisations. AI is the next tool and must be leveraged. What do you think?

Our third keynote speaker, Christopher D. Manning, is giving his talk now. Conference participants can also access the talk remotely. #EMNLP2023

Excited to be at #EMNLP2023! HMU, if you are interested in discussing 🔹 Detecting, attributing & characterizing LLM text 🔹 Reducing LLM hallucinations 🔹 Tackling hate speech & stress on social media Excited to connect & collaborate! 🤝 #LLM #aigeneratedcontent #NLP4SG #EMNLP

This paper delves into the effectiveness and reliability of AI-generated-text detectors, a topic that is increasingly pertinent in today's AI-driven world. You can check out the paper here: lnkd.in/g8mzzkbu #EMNLP2023 #Singapore

I will be attending #EMNLP2023 in #Singapore from December 7th to 10th, and I am excited to share our paper "How Reliable Are AI-Generated-Text Detectors? An Assessment Framework Using Evasive Soft Prompts".

Hi Everyone we are excited to share our latest work on review of augmentation methods to reduce hallucinations in LLMs using Knowledge Graphs. Title: Can Knowledge Graphs Reduce Hallucinations in LLMs? : A Survey @TharinduKTS arxiv.org/pdf/2311.07914… #LLM #hallucinations #KG

🎉 Outstanding Paper! The 'ConDA: Contrastive Domain Adaptation for AI-generated Text Detection' paper by Amrita Bhattacharjee, Tharindu Kumarage, Raha Moraffah and Huan Liu wins the Outstanding Paper Award at #AACL2023! afnlp.org/conferences/ij…

#GoHomeGota2022 #GoHomeRajapakshas

What you cannot create, you do not understand.

@janakact Nayagara 🤩

@janakact 🙋‍♂️