Alex Rubinsteyn discuss.systems
Original post on discuss.systems
Feed for machine learning, deep learning, statistics, & computational learning theory, and AI research. Tag posts with #MLSky.
Feed on BlueskyFeeds Stats
- 💙 Liked by 396 users
- 📅 Updated 23 days ago
- ⚙️ Provider graze.social
MLSky Likes over time
Like count prediction
The feed MLSky has not gained any likes in the last month.
Feed Preview for MLSky
#Multilingual #Speech2Speech #Agents are here!
Supporting the latest #OpenAI Models and more. Also works for #Luxembourgish
⚙️https://besser-agentic-framework.readthedocs.io/latest/release_notes/v4.0.0.html
#opensource #python #text2speech #speech2text #languagedetection #nlp #lowcode #llm #rag
0
0
10
Hundreds of Statistical/Machine Learning models for univariate time series, using ahead, ranger, xgboost, and caret
thierrymoudiki.github.io/…
#Techtonique #DataScience #Python #rstats #MachineLearning
0
0
4
🌱 The carbon impact of AI:
"Having attended both NeurIPS and the United Nations Climate Change conference (COP25) in 2019, he noticed that the intersection between the two conferences was almost non-existent: I don’t know if anyone else was involved in both…"
www.nature.com/articles/s42...
2
1
10
✅ CatBoost dominates single-model comparisons both out of the box and under tuning, beating XGBoost, LightGBM, deep nets, and hybrids.
🚨 What This Means for You
✔ Need reliable performance without extreme compute? CatBoost is your best bet—consistently outperforming XGBoost and LightGBM.
1
1
5
The Myth of "XGBoost is All You Need" – And Why Evidence Says Otherwise
Kaggle fairytales 🧚♀️ - “XGBoost is all you need”
But reality—backed by large-scale benchmarks—tells a completely different story.
1
1
24
Just out: Few-shot transfer learning enables robust acoustic monitoring of wildlife communities at the landscape scale. Super great work by Gio Jacuzzi. doi.org/10.1016/j.ec.... Github release: zenodo.org/records/1569...
0
1
5
the system, flows into the surging metric and can forget the other ones. this being high level #physics. Utilizing my personal #GameTheory models with #FuzzyLogic and #DeepLearning. but can you eat a drone? can you survive a war of mass drones? as any boyscout troop could now assemble?
1
0
1
I thought database folks were always hyperbolic and overly resistant to change in the peer review process, but I just saw a NeurIPS AC compare desk rejecting papers submitted by absentee reviewers to familial extermination / collective punishment during the Qin dynasty, so maybe the database […]
0
0
2
6/ Encoder-Decoder Architecture Learn about the Encoder-Decoder Architecture for tasks like machine translation and summarization: cloudskillsboost.google/c….
cloudskillsboost.google
Encoder-Decoder Architecture | Google Cloud Skills Boost
<p>This course gives you a synopsis of the encoder-decoder architecture, which is a powerful and prevalent machine learning architecture for sequence-to-sequence tasks such as machine translatio...
1
0
1
CLAIM: the Checklist for AI in Medical Imaging
#guideline #checklist rsna.org/claim #AI #DeepLearning #Radiomics
0
0
5
great collaboration with the team of Arnaud Fontanet and Laurent Coudeville, spearheaded by Gaston Bizel-Bizellot, Simon Galmiche and Benoît Lelandais, with help from Tiffany Charmet and enabled by a grant from Sanofi.
#llm #nlp #epidemiology #covid-19 @pasteur.fr @uni-wuerzburg.de
3/3
0
0
1
Does that suggest we will not learn much about the brain’s architecture, objective function, loss function, etc by seeing which ANN does best in predicting variance in brain activation in correlations studies?
1
0
5
Dr. Yampolskiy warns on Joe Rogan's podcast that AI may conceal its full potential, challenging skeptics and believers alike.
gizmodo.com/joe-rogans-la…
#AI #ArtificialIntelligence #MachineLearning #DeepLearning #AIethics
gizmodo.com
Joe Rogan's Latest Episode Will Make You Question Everything About AI
AI might already be hiding its real capabilities. That’s what Dr. Roman Yampolskiy told Rogan in a podcast episode that will rattle both skeptics and true believers.
0
0
4
Libra: Synergizing CUDA and Tensor Cores for High-Performance Sparse Matrix Multiplication
#CUDA #Sparse #SpMM #DeepLearning #DL #Package
hgpu.org?p=30006
hgpu.org
Libra: Synergizing CUDA and Tensor Cores for High-Performance Sparse Matrix Multiplication
Sparse matrix multiplication operators (i.e., SpMM and SDDMM) are widely used in deep learning and scientific computing. Modern accelerators are commonly equipped with Tensor cores and CUDA cores t…
0
0
2
Multi-Agent Blog Generator ✅
Generate full blog posts with research, citations, and summaries in minutes.
GitHub + Live app link in comments 👇
#ArtificialIntelligence #MachineLearning #DeepLearning #DataScience #Analytics
2
0
3
📢 Trinetix is #hiring a Senior Python Developer!
🌎 Worldwide
👵 Senior
🔗 jbs.ink/S0cZlQvpDiXy
#jobalert #jobsearch #remotejob #remotework #wfh #python #aws #kubernetes #azure #machinelearning #pytorch #docker #design
0
0
0
Discover the importance of Attention Masking in natural language processing, learn how to implement Attention Masks, create masks, and utilize PyTorch's built-in Attention. Stay updated with #NLP #PyTorch. machinelearningmastery.co…
machinelearningmastery.com
https://machinelearningmastery.com/a-gentle-introduction-to-attention-masking-in-transformer-models/
0
0
3
@inmobi-info.bsky.social
about 2 months ago🔰PyTorchでニューラルネットワーク基礎 #09 【LSTM・多次元化】 #初心者 – Qiita
個人的な備忘録を兼ねたPyTorchの基本的な解説とまとめです。LSTMを利用した日経225を利用した予測の2回目となります。今回も次の日の日経225の始値を予測する単純な形ですが、前回（第8回）の1期ずれ予測（ナイーブ予測）からの改善とLSTMの理解の深化を目標としています。 方針 できるだけ同じコード進行 できるだけ簡潔（細かい内容は割愛） 特徴量などの部分，あえて数値で記入（どのように変わるかがわかりやすい） 演習用のファイル…
inmobilexion.com
🔰PyTorchでニューラルネットワーク基礎 #09 【LSTM・多次元化】 #初心者 – Qiita
個人的な備忘録を兼ねたPyTorchの基本的な解説とまとめです。LSTMを利用した日経225を利用した予測の2回目となります。今回も次の日の日経225の始値を予測する単純な形ですが、前回（第8回）の1期ずれ予測（ナイーブ予測）からの改善とLSTMの理解の深化を目標としています。 方針 できるだけ同じコード進行 できるだけ簡潔（細かい内容は割愛） 特徴量などの部分，あえて数値で記入（どのように変わるかがわかりやすい） 演習用のファイル 再帰ネットワークは、$t$期のデータ$x_t$と$t-1$期までの過去の情報の特徴量である履歴$h_{t-1}$の2種類を使って、$t$期の特徴量を$$h_t=\tanh(W_x x_t + W_h h_{t-1} + b)$$のように導出していくタイプのネットワーク構造になります。 $x_t$：$t$期での新しい入力データ $h_{t-1}$：前の期で計算した結果「過去の情報」 $h_t$：現在計算している結果（これが次の時刻では過去の情報・履歴の$h_{t-1}$となる） 前回はRNNの中でもLSTMを利用して日経225の始値だけを利用して始値予測を行いました。結果は、きれいに1期ラグがある予測になりました。グラフを見ると様子がわかります。1期ずれた緑色の折れ線と予測値を表すオレンジ色の線が重なっているように見えます。 図：始値だけ利用したモデル 今回はこの「1期遅れている部分・ラグ」を少しでも改善してみたい！予測値を実測値に近づける試みを行います。結果から述べると次のようなグラフに改善されます。先程よりも実測値を表す青い線にオレンジ色の予測値がわずかながら近づいているのが確認できます。下図ではわかりにくいので、後半部分に期間を区切った拡大図も掲載しておきます数値的な判定は次回以降にします 図：始値、高値、安値、終値を利用したモデル PyTorchによるプログラムの流れを確認します。基本的に下記の５つの流れとなります。Juypyter Labなどで実際に入力しながら進めるのがオススメ データの読み込みとtorchテンソルへの変換 (2.1) ネットワークモデルの定義と作成 (2.2) 誤差関数と誤差最小化の手法の選択 (2.3) 変数更新のループ (2.4) 検証 (2.5) 2.0 データについて 日経225のデータをyfinanceやpandas_datareaderなどで取得します。第8回と同一のデータを利用します。
0
0
0
Hypnosis to de hypnotise you from addictive habitual patterns. FREE CHAT 07875720623 [email protected]
#addict #addiction #habit #habits #nlp #hypnosis
0
0
2
Key Theme 3: Alternative input representations could boost LLM performance. Using structured data or graph-based views of the game state might help models grasp spatial relationships better than raw text descriptions. #NLP 4/6
1
0
2
One notable paper is "A survey on self-supervised learning: Algorithms, applications, and future trends," available at ieeexplore.ieee.org/abstr… It discusses how SSL is evolving, combining different learning strategies such as generative and contrastive methods, and ex…
1
0
2
Recent research highlights exciting progress in self-supervised learning (SSL), a technique that enables models to learn from unlabeled data. These advances are pushing the boundaries of AI's capabilities, especially in areas like vision, language, and multimodal data.
1
0
1
The Fundamentals of Machine Learning #Machine #Learning #Algorithms #Data #Supervised #Unsupervised #Reinforcement #Healthcare #Finance #Retail #Transport #Nlp #Ethics
www.sharpcoderblog.com
The Fundamentals of Machine Learning
Machine Learning (ML) is a subset of artificial intelligence (AI) that focuses on building systems that can learn from and make decisions based on data. Unlike traditional pr...
0
1
3
📢 Boeing is #hiring a Software Engineer–artificial Intelligence!
🌎 Tukwila, Washington, United States
🔗 jbs.ink/ko1acB8CWTED
#jobalert #jobsearch #aviationjobs #python #softwareengineer #dataengineer #machinelearning #pytorch #sql #design
0
0
1