You're in an ML Engineer interview at Apple. The interviewer asks: "Two models are 88% accurate. - Model A is 89% confident. - Model B is 99% confident. Which one would you pick?" You: "Any would work since both have same accuracy." Interview over. Here's what you missed: Modern neural networks can be misleading. They are overconfident in their predictions. For instance, I saw an experiment that used the CIFAR-100 dataset to compare LeNet with ResNet. LeNet produced: - Accuracy = ~0.55 - Average confidence = ~0.54 ResNet produced: - Accuracy = ~0.7 - Average confidence = ~0.9 Despite being more accurate, the ResNet model is overconfident in its predictions. While the model thinks it's 90% confident in its predictions, in reality, it only turns out to be 70% accurate. Calibration solves this. A model is calibrated if the predicted probabilities align with the actual outcomes. For instance, say a model predicts an event with a 70% probability. Then, ideally, out of 100 such predictions, ~70 should result in the event. Handling this is important because the model will be used in decision-making. In fact, an overly confident that is not equally accurate model can be highly misleading. To exemplify, say a government hospital wants to conduct an expensive medical test on patients. To ensure that the govt. funding is used optimally, a reliable probability estimate can help the doctors make this decision. If the model isn't calibrated, it will produce overly confident predictions. Reliability Diagrams are a visual way to inspect how well the model is currently calibrated. More specifically, this diagram plots the expected sample accuracy as a function of the corresponding confidence value (softmax) output by the model. If the model is perfectly calibrated, then the diagram should look like the identity function. That said, it is often also useful to compute a scalar value that measures the amount of miscalibration, called expected calibration error (ECE). One way to approximate the expected calibration error shown above is by partitioning predictions into equally spaced bins and taking a weighted average of the bins’ accuracy/confidence difference. These are some common techniques to calibrate ML models: > For binary classification models: - Histogram binning - Isotonic regression - Platt scaling > For multiclass classification models: - Binning methods - Matrix and vector scaling 👉 If you care about probabilities and both models are operationally similar, which model would you prefer? ____ Find me → @_avichawla Every day, I share tutorials and insights on DS, ML, LLMs, and RAGs.

DSperse is now powering ML workloads on Subnet-2. Slice models → prove parts → scale what used to be impossible. This is what production zkML infrastructure actually looks like.

Why did the report evaluate ML-DSA-44 specifically? The biggest reason is practicality. Compared to higher-security parameter sets like ML-DSA-65 and ML-DSA-87, ML-DSA-44 offers: • Smaller signatures • Faster verification • Lower network overhead That matters because signature size quickly becomes the dominant scaling constraint in post-quantum systems. In testing, PQ blocks were already ~18× larger than non-PQ blocks at equivalent TPS. Read the report 👇

200 of the sharpest AI and ML engineers will go from idea to working product in 36 hours at the NEXT hackathon. US$130,000 in prizes, backed by @awscloud and @vercel . This is the room you want to be in if you ship. Applications close 19 May:

BNB chain just published concrete post-quantum migration research using ML-DSA-44 signatures and pqSTARK aggregation. bitcoin has zero BIPs for quantum resistance. ethereum has zero EIPs with timelines. google/stanford research compressed the quantum threat from 20-30 years to 5-10 years and the market is pricing exactly 0% premium for chains that are actually preparing. BNB's 21 validators can coordinate a cryptographic hard fork in months. ethereum's 900,000 validators took 7 years to ship the merge. when IBM hits 50k stable qubits around 2029, the coordination problem becomes an existential problem. the "centralization discount" on BNB flips into a survival premium during forced migration. still early

I’m actively looking for AI Engineer | ML Engineer | Forward Deployed Engineer roles (India or USA) 🇮🇳🇺🇸 Just finished at JP Morgan Chase building production LLM agentic AI workflows: 👉 GPT-4 compliance chatbot with LangChain + LangGraph ➡️ cut manual review time by 30% 👉 Real-time fraud detection systems (PySpark + Azure OpenAI) ➡️ reduced false positives by 15% Specialist in RAG, LLM fine-tuning, and scalable MLOps pipelines. Open to work immediately. Let’s connect! 🔥

See the top ranked papers in AI, ML, Robotics, Quantum Physics, and more on @kurateorg. Hundreds of arXiv preprints ranked daily by scientific impact through pairwise tournaments judged by Claude, GPT, and Gemini.

We’ve published a technical report evaluating a post-quantum cryptography migration path for BSC. The report covers: • ML-DSA-44 transaction signatures • pqSTARK validator aggregation • Type 0x05 transaction format • Public key storage and verification flow • Cross-region performance benchmarks Benchmark highlights: • Tx size: 110 B → ~2.5 KB • Block size: ~110 KB → ~2 MB • Native transfer TPS: 4,973 → 2,997 One notable result: The primary bottleneck was not signature verification performance, but block byte size and cross-region propagation overhead. Read the full report 👇