Пікірлер
@radicalengineer2331
@radicalengineer2331 3 күн бұрын
Are you taking student this fall, Prof?
@mobinkarimaei7865
@mobinkarimaei7865 6 күн бұрын
You are the best. I wish you health and happiness, dear Dr. Karimi.🙏
@amirhkarimi
@amirhkarimi 6 күн бұрын
I’m glad you enjoyed the pod 🙏🏼
@getiyeabule6260
@getiyeabule6260 8 ай бұрын
can you share me the ppt
@amirrezamohammadi
@amirrezamohammadi Жыл бұрын
Nice and Coherent talk, Specially enjoyed the idea to link the concepts and program synthesis to XAI. That would be the future of the XAI considering the LLM breakthroughs. Thanks Amir-Hossein Jan :)
@amirhosseinahmadi6759
@amirhosseinahmadi6759 Жыл бұрын
Too much for a first year data science student at Laurier to understand! Shine on Amir 💫
@amirhkarimi
@amirhkarimi Жыл бұрын
Thanks Amirhossein🙏🏼🙏🏼 same to you, keep rocking 💪🏼
@LongTran-tr8sx
@LongTran-tr8sx 2 жыл бұрын
great video, thanks for your contribution!
@DistortedV12
@DistortedV12 3 жыл бұрын
volume too low
@amirhkarimi
@amirhkarimi Жыл бұрын
🫠
@bingyangwen958
@bingyangwen958 4 жыл бұрын
Great talk! Can I ask a couple of questions? Do you think the method of counterfactual explanations implicitly assume the causal relation: X->Y, where X is input variables and Y is the model's output.
@amirhkarimi
@amirhkarimi 4 жыл бұрын
Indeed, from a temporal perspective, we may consider that X's precede Y and are therefore a cause (at least when the model h is involved). However, we don't rely on this relationship when deriving the structural counterfactual, we only use h() to solve the optimization problem. Does this help answer you question?
@bingyangwen958
@bingyangwen958 4 жыл бұрын
@@amirhkarimi Yes, thanks for your explanation!
@علیرضادربندی-ث8ب
@علیرضادربندی-ث8ب 4 жыл бұрын
آفرین
@Cipherweave
@Cipherweave 4 жыл бұрын
its sooo beautiful. Im glad having a brother like you :)
@Cipherweave
@Cipherweave 4 жыл бұрын
oooooooooffffffffffffffffffffffffffffffff
@florianro.9185
@florianro.9185 5 жыл бұрын
Very nice... Are there any other papers you can suggest for counterfactual explanations? I already looked at CertifAI, MACE, FACE, DiCE, Foil Trees and CLEAR. Thanks
@amirhkarimi
@amirhkarimi 5 жыл бұрын
Those are all great options! Perhaps you would find these interesting as well: - arxiv.org/abs/1706.06691 - arxiv.org/abs/1711.00399 - arxiv.org/abs/1712.08443 - arxiv.org/abs/1809.06514 - arxiv.org/abs/1907.09615 - arxiv.org/abs/1910.00057
@fs.azimifar
@fs.azimifar 5 жыл бұрын
I'm proud of you ❤️