Відео

Very Good - I have been in coding and reporting for many decades - and you are quite right. One of the most successful companies I worked for produced what I thought were very basic reports. Bog simple - but the customers loved them. Technically simple but the owner knew her customers so well and the reports had nothing to do with us or the tools capabilities - Beauty is in the eye of the beholder - or your customer

Thank you guys, really enjoyed this!

It would be nice to have a link to the slides....maybe with some way to link comment threads to each. I love3 the concept of NLP to the query language. I hate having to learn varying syntax for differing data structures implemented in competing products....I would love to have international standards for all common commands with knowledge bases which could each have linked (graphed) annotations for ongoing issues.

how is entity discovery done? Are you doing Unsupervised clustering in the embedding space? and then labeling the clusters?

nice!

Great presentation! Are the slides available somewhere?

The problem ai see is each company defining its standard. Isn’t there already a standard for defining data contracts that organizations already use ? How does that relate to dcat? And Dublin core?

Combining LLMs with decentralized/distributed knowledge graphs would be even better because it provides for greater optionality and prevents vendor lock in. Projects like Origin Trail are already doing this at scale with clients like the British Standards Institute.

Great talk but one of the fundamental problems of neo4j is it's lack of scalability to actually handle large data. When your database grows into the hundreds of millions of nodes and relationships it becomes a nightmare to work in

a tutoriel of how this done would be amazing

One of the worst Softwares i came across. Poor Documentation, hard to setup and breaking updates all the time

wow, this does get me excited about graph and LLM!

This was a fun experience - thanks Big Data LDN :)

Great video on Data Strategy!! Highly recommend to anyone who's building enterprise level data strategy for your organization!

Nice, this helps me !

LOL, "what king" magazine 🤣. And python snake oil. Genius.

Loving this, what an engaging start, and bang on!

The argument made by *some* doomers is not that we are close to AGI, but that when it eventually comes, we are doomed, unless we can figure out a way to align it. The remaining distance may be great in terms of capability, yet short in terms of time if we keep accelerating the development.

Ah, dude, free Y42 t-shirt?! Send me one!

Video Summary: 结合大型语言模型和知识图谱，以增强知识表示。 - 00:00 这部分视频介绍了知识图谱的定义以及演示了如何将大型语言模型与知识图谱相结合。 - 04:10 在大型语言模型中，语义概念和图谱是非常重要的，其中有两个关键领域，一个是利用大型语言模型来创建机器可读的语义，另一个是通过图机器学习从网络结构中学习事实信息。 - 08:23 通过将知识图谱与大型语言模型相结合，可以提取子图并进行进一步分析，以获得更好的知识表示。 - 12:35 这一部分介绍了如何将知识图谱与大型语言模型结合起来增强知识表示。 - 16:47 为了解决这个问题，我们可以使用知识图谱来强制大型语言模型使用特定的数据集，这个数据集基于你提供的事实知识，从而减少模型的虚构性。 - 21:00 通过使用知识图谱进行基于大型语言模型的知识表示的优势 - 25:11 可以通过将文章转化为向量并进行语义搜索来增强知识表示。

Nice! Great job Ammara 🙌

Is it true data mesh ?

14:00 "Kimball or Inmon or...."

4:20 "Invented" might be too strong a word. :)

3:35 Kimball

Wonderful presentation. Thank you.

Great presentation! Congratulations! Well done!!! 👍👍👍🙏🙏🙏

Very informative! ChatGPT’s responses are concise too, so i can see how it’s cost-effective

Thanks Neeraj Goel Mittal for explaining in so simple way.. Keep it up

When I joined digital fineprint i was astonished to find the dev and data teams not only separate, but they didnt even talk to each other, it was madness. Very quickly i joined those teams together and we operated as one - and now we do exactly the same at Taveo.

can i get the git repo

good design sharpens the attention of the viewer and raises his/her level of attention. you do not just consume the information, you work with it.

Hello, I am working in deliveroo as a riderI am just interested what programs or systems do you use for planning or forecasting the best route for riders and drivers to pick orders? Might i would be interesting to see and to have practice to analyze? As well do you have plan or forecast or past data when you on shortage of riders and in which areas or cities?

The negatives with Apache Pinot as I understand it is: slow performance (4x slower on ClickHouse Benchmark compared to StarRocks), no Cost Based Optimizer (which means that your queries may not get faster over time) and cannot query Apache Iceberg or Apache Hudi or any of the open table formats on data lake (can't take advantage of open analytics formats and cloud based storage).

How to do the Report Navigation Help?

Very insightful& valuable information about Data Fabric here, thanks for sharing!

audio is really bad on this - just needs to filter out the background noise.

Ben zo fier op je mijn zoon

That was awesome!

Tower of Babel is the great metaphor for lack of interoperability.

Great interview. SUPERB, well noted.

Love the "about me" slide :)

Beta tester od roku 2016 oceňujem,,

Brilliant talk Sarah!

❤

Gald to hear you, Tina. Very impressive.

Thank you so much Big Data London for having me again this year.

Pointless, absolutely "Pointless" ua-cam.com/video/MxX6b9cRyaw/v-deo.html.. . . . Yes the Pointless TV show (and quiz book) provides some great examples of all those biases of how bad (or good) the sample data can be, how our selection algorithms can go horribly wrong, and how being in some minority group (e.g. lives in Wales, knows Welsh stuff; lives in Scotland, knows Scottish stuff) can completely change the outcomes. Sorry about the click bait first line. Just been out to dinner at friends and played a few rounds of "Pointless", and it struck me how much it was an example for the ML/AI issues.

Algorithm biases: thoughts - some expect long (fat) tails, some look for central shape, some look for the steepness of slope to determine transitions, etc. While these are also data artefacts, they are predominately algorithmic - if you favour one algorithmic type then you get the same failures. E.g. logistic regression vs Gaussian fits - the logistic has an almost identical shape for the matched distribution but 'ignores' (discriminates against) [potentially important] outliers. Data normalisation enhances some properties (e.g. rotatability of the distribution) favouring certain algorithms, so where you place it in the analysis stream can affect the twitchiness of the result.