small language model

AI Contexts That Never Forget - How to give an artificial intelligence a world it keeps

A frontier model writes working code and solves competition mathematics. Ask the same model to reconcile three hundred invoices and it loses track of which ones it already processed. A spreadsheet has done that task perfectly since 1985. This article locates the difference, measures it, and names the one operation that...

Adding Noise Can Improve Detection ── But "More Is Better" Is Not the Rule ── Zero noise gives detection rate 0, sigma=0.50 gives correlation 0.687458, sigma=4 gives 0.127152 ── The optimum is not at either end but in the middle ── [Paper 353]

One is taught that noise is to be reduced. This paper asks whether reducing it always improves matters──the answer is that in a system with a threshold it does not. No new mathematical theorem and no new law is claimed. Scope of this paper (scope note): No new mathematical theorem and no new law is claimed──stochastic...

The "Depth" of Depth of Field Is a Convention, Not a Measurement ── With the same 50 mm lens at the same f/2.8, the hyperfocal distance is 29.8119 m and also 207.6912 m ── What changed is not the equipment but the circle of confusion, which is an agreement ── One set of equipment holds four answers ── [Paper 358]

Depth of field is printed in tables and looks like a property of the lens. This paper asks where that number comes from──the answer is that it comes not from the equipment but from an agreement about when blur becomes visible. No new mathematical theorem and no new law is claimed. Scope of this paper (scope note): No n...

Pilot comparison of nine large language models generating radiological reports from dental panoramic radiograph pathology notation

Large language models (LLMs) may help convert structured pathology notations derived from dental panoramic radiographs (DPRs) into radiological reports, but comparative data remain limited. This pilot head-to-head study evaluated nine publicly available LLMs across five domains: correct tooth numbering, correct diagn...

Two Algebraically Equal Forms Return Answers Differing by 15 Digits ── The smaller root of a quadratic comes out with 0.59 correct digits from the naive form and 16 from the rationalised one ── The cause is that sqrt(b^2-4c) agrees with b to 15.83 digits ── Subtract two near-equals and exactly the agreeing digits vanish ── [Paper 368]

It is tempting to think that mathematically identical formulae return identical answers. This paper asks whether that holds in finite precision──the answer is that it does not, and the gap reaches 15 digits. No new mathematical theorem and no new law is claimed. Scope of this paper (scope note): No new mathematical the...

ShalDeepDP: A Survey and Comparison of Data Poisoning in Statistical, Classical Deep Learning and Foundation Models

Advances in real-world adversarial threats have heightened concerns over the privacy and security of AI systems. One key threat is Data Poisoning, where malicious data is injected into training sets or model inputs to compromise model behavior. While much of the current research focuses on deep learning, particularly F...

AI Contexts That Never Forget - How to give an artificial intelligence a world it keeps

A frontier model writes working code and solves competition mathematics. Ask the same model to reconcile three hundred invoices and it loses track of which ones it already processed. A spreadsheet has done that task perfectly since 1985. This article locates the difference, measures it, and names the one operation that...

RSEs as trainers – the who, why and how of providing digital research skills training

RSEs often provide training for researchers and dRTPs at their local institution in the digital research skills used in their own work. These courses cover a diverse range of topics, from beginners' language courses, to software architecture and efficiency, to specialised tools and platforms like HPC, AI and domain-spe...

In RAG We Trust? Measuring Robustness of Retrieval-Augmented Generation Under Document Poisoning

Retrieval-augmented generation (RAG) grounds a language model in retrieved documents, which reduces hallucination but creates a new attack surface: if retrieved text is tampered with, the model may repeat the falsehood. We study how much a small quantized model, Llama 3.1 8B, degrades when a fraction of its retrieved c...

Tact: A Zero-Cost, Browser-Based Pipeline for On-Demand Tactile Braille Storybooks

Tact, a browser-based pipeline that converts a spoken or typed story idea into printable braille with a matching raised tactile illustration, without an account or mandatory cost and with an offline-capable path is presented.

Tact: A Zero-Cost, Browser-Based Pipeline for On-Demand Tactile Braille Storybooks

Braille literacy among blind school-age children fell from roughly 50 percent in 1960 to under 10 percent today, and the decline is a content problem, not a technology problem: producing one illustrated Braille page still requires specialized software and trained labor, so almost nothing in the existing catalog is pers...

Exploring the Dynamic Interplay Between Intrinsic Value, Perceived Costs, and Language Achievement: A Cross-Lagged Panel Analysis