Euclidean distance is sensitive to the magnitudes of the vectors, whereas cosine similarity is not.

Linear models with a rank constraint and a fast loss approximation can train on a billion words within ten minutes, while achieving performance on par with the state-of-the-art for tag prediction and sentiment
analysis.
#aigraph

Sentence transformers are LLMs that are specially trained to output embeddings at the sentence level rather than the word or token level. These sentence level embeddings do a better job of capturing meaning and are far quicker to compute.
#aigraph

CLUSTERLLM is a novel text clustering framework that takes feedback from an instruction-tuned large language model, such as ChatGPT to guide a small embedder for finding text clusters with low cost.
#aigraph

Large Language Models can be leveraged to direct clustering algorithms during semi-supervised learning, thus requiring less input from the experts.
#aigraph

Persistent identifiers (like ORCID and DOI) can maximise the value of the research investments by
1. Improving research quality
2. Catalysing innovation and impact
3. Increasing the efficiency of the research ecosystem
Ref:

#aigraph FastText, just as its name 'fast' suggests, the biggest advantage is that it can often achieve accuracy comparable to deep networks, but its training time is many orders of magnitude faster than deep networks.

fastText can be trained on more than one billion words in less than ten minutes using a standard multicore~CPU, and classify half a million sentences among~312K classes in less than a minute.
#aigraph