Saturday, June 28, 2025
Social icon element need JNews Essential plugin to be activated.
No Result
View All Result
Crypto now 24
  • HOME
  • BITCOIN
  • CRYPTO UPDATES
    • GENERAL
    • ALTCOINS
    • ETHEREUM
    • CRYPTO EXCHANGES
    • CRYPTO MINING
  • BLOCKCHAIN
  • NFT
  • DEFI
  • METAVERSE
  • WEB3
  • REGULATIONS
  • SCAMS
  • ANALYSIS
  • VIDEOS
MARKETCAP
  • HOME
  • BITCOIN
  • CRYPTO UPDATES
    • GENERAL
    • ALTCOINS
    • ETHEREUM
    • CRYPTO EXCHANGES
    • CRYPTO MINING
  • BLOCKCHAIN
  • NFT
  • DEFI
  • METAVERSE
  • WEB3
  • REGULATIONS
  • SCAMS
  • ANALYSIS
  • VIDEOS
No Result
View All Result
Crypto now 24
No Result
View All Result

LLaVA vs. GPT-4: An Open-Source AI Showdown Highlighting Multimodal Potential and Mathematical Limitations

September 6, 2023
in Metaverse
Reading Time: 4 mins read
A A
0

[ad_1]

On the current presentation of GPT-4, one of many standout options was its capability to have interaction in conversations enriched with photos. Nonetheless, this functionality is but to be built-in into OpenAI’s providing. Whereas we beforehand highlighted Bing’s competence on this side, an open-source answer has now emerged within the type of the “Giant Language and Imaginative and prescient Assistant” (LLaVA).

Mathematical Challenges: LLaVA's Open Source AI Struggles, Unlike GPT-4

Printed: 6 September 2023, 4:27 pm Up to date: 06 Sep 2023, 4:28 pm

LLaVA’s Multimodal Potential

LLaVA is an open-source multimodal AI that mixes language and imaginative and prescient processing. A demo of LLaVA could be accessed right here.

The odd factor about this image is {that a} man is holding a clothesline with a white shirt on it as he leans out of the window of a yellow taxi. This case is uncommon as a result of it’s unusual to see individuals leaning out of automotive home windows whereas holding clothes or undergarments. The person might be making an attempt to dry his shirt in an unsafe and inefficient method whereas the taxi is shifting.

We carried out a easy check by inputting an image of Taxi and man, and LLaVA supplied me with a descriptive evaluation. Nonetheless, my makes an attempt to problem it with mathematical issues, much like these tackled by Bing, proved futile. LLaVA seems to wrestle with arithmetic, regardless of its proficiency in picture recognition.

Challenges with Mathematical Duties

We introduced LLaVA with a mathematical drawback involving trigonometry, akin to these efficiently solved by Bing. Sadly, LLaVA couldn’t present the right answer, returning an faulty reply for the sq. root of 169 as 13.2.

Let’s resolve this question

The duty at hand entails figuring out the size of the diagonal of a triangle with sides measuring 5 cm and 12 cm. To perform this, we will make use of the Pythagorean theorem, which states that in a right-angled triangle, the sq. of the hypotenuse’s size (the aspect reverse the correct angle) equals the sum of the squares of the opposite two sides.

On this state of affairs, the hypotenuse corresponds to the diagonal of the triangle, which we will confirm by taking the sq. root of the sum of the squares of the 2 given sides.

To start, we should discover the size of the diagonal, which is the aspect connecting the 2 non-adjacent vertices of the triangle. On this occasion, these two vertices are related to the perimeters measuring 5 cm and 12 cm.

Now, let’s apply the Pythagorean theorem:

(5 cm)^2 + (12 cm)^2 = diagonal^2

5^2 + 12^2 = diagonal^2

25 + 144 = diagonal^2

169 = diagonal^2

Lastly, we will decide the size of the diagonal by taking the sq. root of either side:

√169 = diagonal

The size of the diagonal is roughly 13.2 cm.

An Uncommon Perspective on Photographs

LLaVA excels in its capability to converse with photos, though challenges persist, notably in mathematical problem-solving. For instance, when introduced with a picture of a person leaning out of a yellow taxi window, holding a clothesline with a white shirt, LLaVA supplied an uncommon perspective. It steered that such a scene is atypical, as it’s not widespread to witness people leaning out of automotive home windows whereas holding clothes. The evaluation indicated that the person could also be making an attempt an unconventional and doubtlessly unsafe methodology of drying his shirt whereas the taxi is in movement.

Whereas LLaVA provides promising multimodal capabilities, notably in conversing with photos, it faces limitations in mathematical problem-solving. It’s price noting that Google’s capabilities on this regard surpass LLaVA’s, as demonstrated by a extra correct answer to an analogous mathematical drawback.

The event of AI with multimodal capabilities is undoubtedly an thrilling development, and LLaVA is a commendable open-source effort on this route. Nonetheless, enhancements are wanted to boost its mathematical reasoning capabilities to match its proficiency in picture evaluation.

For a extra correct mathematical answer, Google’s capabilities are at the moment superior: Google’s Mathematical Downside Solver.

Learn extra about AI:

[ad_2]

Source link

Tags: GPT4highlightingLimitationsLLaVAMathematicalMultimodalOpenSourcepotentialShowdown
Previous Post

Ghostwriter Drops AI Travis Scott Song, Aims for a GRAMMY

Next Post

Renzo Piano to design performing-arts centre in South Florida

Next Post
Renzo Piano to design performing-arts centre in South Florida

Renzo Piano to design performing-arts centre in South Florida

Blackrock’s Bitcoin Spot ETF May Unleash $30 Trillion From US Advisors

Blackrock's Bitcoin Spot ETF May Unleash $30 Trillion From US Advisors

SNXweave Weekly Recap 105

SNXweave Weekly Recap 105

Leave a Reply Cancel reply

Your email address will not be published. Required fields are marked *

Social icon element need JNews Essential plugin to be activated.

CATEGORIES

  • Altcoin
  • Analysis
  • Bitcoin
  • Blockchain
  • Crypto Exchanges
  • Crypto Mining
  • Crypto Updates
  • DeFi
  • Ethereum
  • Metaverse
  • NFT
  • Regulations
  • Scam Alert
  • Uncategorized
  • Videos
  • Web3

SITE MAP

  • Disclaimer
  • Privacy Policy
  • DMCA
  • Cookie Privacy Policy
  • Terms and Conditions
  • Contact us

Copyright © 2023 Crypto Now 24.
Crypto Now 24 is not responsible for the content of external sites.

No Result
View All Result
  • HOME
  • BITCOIN
  • CRYPTO UPDATES
    • GENERAL
    • ALTCOINS
    • ETHEREUM
    • CRYPTO EXCHANGES
    • CRYPTO MINING
  • BLOCKCHAIN
  • NFT
  • DEFI
  • METAVERSE
  • WEB3
  • REGULATIONS
  • SCAMS
  • ANALYSIS
  • VIDEOS

Copyright © 2023 Crypto Now 24.
Crypto Now 24 is not responsible for the content of external sites.