text stringlengths 1 1k ⌀ | title stringclasses 230
values |
|---|---|
At this stage, I found the second explanation probably makes more sense. Language models
are designed to generate readable texts. They do not have a deep "understanding" of the
physical world, nor are they trained to do sophisticated reasoning. Think about how we get
to understanding the world — reading newspapers and ... | OpenAI's GPT-3 Language Model_ A Technical Overview |
6 SUMMARY AND LIMITATIONS
We have presented GPTQ, an approximate second-order method for quantizing truly large language
models. GPTQ can accurately compress some of the largest publicly-available models down to 3
and 4 bits, which leads to significant usability improvements, and to end-to-end speedups, at low
accuracy... | GPTQ |
3.2 COLLISIONS | MULTI HASH EMBEDDINGS IN SPACY |
and frame-level mask m can be written as m = rep(m′, l). For the applications where a duration
model is involved (zero-shot TTS, content editing, diverse speech sampling), the frame-level mask
m is extended such that no phone is partially masked. In other words, all the frames corresponding
to a phone is either entirel... | Voicebox-Text-GuidedMultilingual UniversalSpeechGenerationatScale |
A. ACCURACY
The accuracy score, also known as the classification accuracy
rating, is determined as the percentage of accurate predictions
in proportion to the total predictions made by the model.
The accuracy (A) can be depicted by the given formula in
Equation (1).
A = TruePositive + TrueNegative
TotalNumberofPredict... | A_Comprehensive_Review_on_Fake_News_Detection_With_Deep_Learning |
media text. In Proceedings of the international AAAI conference on web and social media, volume 8, pages
216–225, 2014.
Mandar Joshi, Eunsol Choi, Daniel S Weld, and Luke Zettlemoyer. Triviaqa: A large scale distantly supervised
challenge dataset for reading comprehension. arXiv preprint arXiv:1705.03551, 2017. | Llama2 |
3 A FROZEN LM READER FOR OPEN-DOMAIN QUESTION ANSWERING | STANDING ON THE SHOULDERS OF GIANT FROZEN LANGUAGE MODELS |
23This does not refer to the downstream risks after deploy-
ment, just the risks associated with the research experiments.
Two reasons for NLP experiments being relatively low risk
are the rare involvement of human participants in NLP exper-
iments and the historical focus on professionally generated
text (Hovy and Spr... | A Two-Sided Discussion of Preregistration of NLP Research |
T
h
e
l
a
y
e
r
s
i
n
t
h
e
m
i
d
d
l
e
c
r
e
a
t
e
s
h
o
r
t
c
u
t
s
t
o
s
p
e
e
d
u
p
s
e
a
r
c
h
.
W
h
e
n
p
e
r
f
o
r
m
i
n
g
a
s
e
a
r
c
h
,
H
N
S
W
s
t
a
r
t
s
f
r
o
m
a
r
a
n
d
o
m
n
o
d
e
i
n
t
h
e
t
o
p
l
a
y
e
r
a
n
d
n
a
v
i
g
a
t
e
s
t
o
w
a
r
d
s
t
h
e
... | LLM Powered Autonomous Agents _ Lil'Log |
Rodriguez, P., Barrow, J., Hoyle, A. M., Lalor, J. P., Jia, R., and Boyd-Graber, J. Evaluation examples are not
equally informative: How should that change NLP leaderboards? In Proceedings of the 59th Annual Meeting of
the Association for Computational Linguistics and the 11th International Joint Conference on Natural ... | PaLM 2 Technical Report |
Impact on Innocent People: Rumors can have a major
impact on specific people. These people may be harassed
by social media. They may also face insults and threats
that may have real-life consequences. People must not
believe in invalid information on social media or judge a
person.
Impact on Health: The number of peopl... | A_Comprehensive_Review_on_Fake_News_Detection_With_Deep_Learning |
In an autoregressive language model, such as GPT-3 and PaLM [22], given a context sequence 𝑋,
the LM tasks aim to predict the next token 𝑦. The model is trained by maximizing the probability
of the given token sequence conditioned on the context, i.e., 𝑃(𝑦|𝑋) = 𝑃(𝑦|𝑥1, 𝑥2, ..., 𝑥𝑡−1), where
𝑥1, 𝑥2, ..., 𝑥... | ASurveyonEvaluationofLargeLanguageModels |
Omkar M. Parkhi, Andrea Vedaldi, Andrew Zisserman, and C. V. Jawahar. Cats and dogs.
In IEEE
Conference on Computer Vision and Pattern Recognition, 2012.
Deepak Pathak, Philipp Krähenbühl, Jeff Donahue, Trevor Darrell, and Alexei Efros. Context encoders:
Feature learning by inpainting. In CVPR, 2016.
David Patterso... | DINOv2- Learning Robust Visual Features without Supervision |
Prompting. Prompting is a way of leveraging a
pretrained LM by prepending instructions and a
few examples to the task input and generating the
task output from the LM. For autoregressive LMs,
the most successful form of prompting is GPT-3’s
in-context learning (Brown et al., 2020), which
uses manually designed prompts ... | Prefix-Tuning |
5.2.1 Setup and Implementation Details
We conduct experiments on both finetuning and in-context learning. For supervised finetuning, our models
are continuously finetuned after N pretraining steps where N is typically from 50k to 100k. In other words,
after each Nk steps of pretraining, we finetune on each downstream task ... | UL2- Unifying Language Learning Paradigms |
task to respond to the input task, and these target contents
serve as the information context for retrieving more relevant
knowledge. This helps to generate better responses in another
iteration. | Retrieval-AugmentedGenerationforLargeLanguageModels-ASurvey |
Ilya Sutskever, and Wojciech Zaremba. Evaluating large language models trained on code. arXiv
preprint arXiv:2107.03374, 2021. URL https://arxiv.org/abs/2107.03374. | gemini_1_report |
1 3 Social Network Analysis and Mining (2021) 11:32
Page 11 of 15 32
highly biased and did not increase their bias ratings after
seeing the flags.) These responses suggest that efforts to
flag bot content may mitigate their ability to spread mis-
information on Twitter. These results align with previous
studies ... | Use of bot and content flags to limit the spread of misinformation among social networks: a behavior and attitude survey |
4.6 Comparison to other methods
We now compare the performance of our final model with competitive baselines: EnCodec [8], Lyra
[46], and Opus [37], a popular open-source audio codec. For EnCodec, Lyra, and Opus, we use
publicly available open-source implementations provided by the authors. We compare using both
objec... | RVQGAN |
2https://gluebenchmark.com/
Multi-headed attentionLayer Norm+Adapter2x Feed-forward layerLayer Norm+AdapterFeed-forward layerTransformerLayerNonlinearityFeedforward up-projectFeedforward down-projectAdapter Layer+ Parameter-Efficient Transfer Learning for NLP
We compare to fine-tuning, the current standard for transfer... | Parameter-Efficient Transfer Learning for NLP |
I’ll say that an AI system has “advanced capabilities” if it outperforms the best humans on some set of
tasks which when performed at advanced levels grant significant power in today’s world. The type of
tasks I have in mind here include: scientific research, engineering, business/military/political strategy,
hacking, an... | Is Power-Seeking AI an Existential Risk? |
Language models can explain neurons in language models
https://openaipublic.blob.core.windows.net/neuron-explainer/paper/index.html
14/32 | Language models can explain neurons in language models |
Vis Comput 2020; 101: 103968.
[55] Chatterjee J and Dethlefs N. XAI4Wind: a multimodal knowledge graph database for explainable decision support in operations
& maintenance of wind turbines, 2020, https://arxiv.org/abs/2012.10489
[56] Shi N, Zeng Q and Lee R. The design and implementation of language learning chatbo... | Knowledge-graph-based explainable AI- A systematic review |
Richard S Sutton and Andrew G Barto. Reinforcement learning: An introduction. MIT press, 2018.
Alon Talmor, Jonathan Herzig, Nicholas Lourie, and Jonathan Berant. Commonsenseqa: A question
answering challenge targeting commonsense knowledge. In Proceedings of the 2019 Conference
of the North American Chapter of the As... | LARGELANGUAGEMODELSCANNOTSELF-CORRECT REASONINGYET |
References
Armen Aghajanyan, Bernie Huang, Candace Ross, Vladimir Karpukhin, Hu Xu, Naman Goyal, Dmytro
Okhonko, Mandar Joshi, Gargi Ghosh, Mike Lewis, and Luke Zettlemoyer. CM3: A causal masked
multimodal model of the internet. arXiv:abs/2201.07520, 2022.
Wasi Uddin Ahmad, Saikat Chakraborty, Baishakhi Ray, and Kai-W... | CodeLlama2 |
8
such as software development [108; 109] and scientific research [110]. Due to their natural language
comprehension and generation capabilities, they can interact with each other seamlessly, giving rise
to collaboration and competition among multiple agents [108; 109; 111; 112]. Furthermore, research
suggests that a... | TheRiseandPotentialofLargeLanguageModel BasedAgents |
1.88
3.12
3.88
0.30
1.67
3.33
4.33
0.41
1.78
3.33
4.44
0.41
1.25
2.75
4.00
0.41
1.60
3.20
4.10
0.43
2.37
3.07
3.57
0.20
2.43
3.50
3.92
0.23
2.12
3.35
4.08
0.28
1.77
2.55
3.60
0.29
2.78
3.30
3.80
0.18
51
Table 11: Table summarizes distributions of LLM-reported personality
scores on the various Psychometric measures co... | PersonalityTraitsinLargeLanguageModels |
P
I
V
L
L
64
62
1
-
d
l
o
F
C
S
E
24
22
20
18
D
4
o
g
E
ViT-B ViT-L
ViT-H
ViT-B ViT-L
ViT-H
Figure 6. Scaling the image encoder size while keeping the other
modality encoders’ size fixed. We measure the performance on
the emergent zero-shot classification of depth, audio, thermal, and
IMU modalities. Scaling ... | IMAGEBIND- One Embedding Space To Bind Them A |
exact wording of their input when deciding whether
or not to call an API; this is perhaps unsurprising
given that LMs are known to be very sensitive to
the prompt they are provided with in both zero-
and few-shot settings (Jiang et al., 2020; Schick
and Schütze, 2021a). Depending on the tool, our
method is also very sa... | Toolformer |
1
Preprint | CRAMMING-TRAININGALANGUAGEMODELONA SINGLEGPUINONEDAY |
disclaimers, which was part of a round of revisions the company made in
response to public pressure in the wake of the 2016 election. These included
a formal verification that ad buyers have US-based residences. We discuss
Facebook’s efforts in more detail in the section on the platform’s response to
the 2016 election. | Social_Media_and_Democracy |
Training Data. To achieve state-of-the-art reconstruction
quality, we collect 1000 high-quality textured human scans
with various clothing, shapes, and poses from Twindom1.
1. https://web.twindom.com/
(a)(b)(c)(d) JOURNAL OF LATEX CLASS FILES, VOL. 14, NO. 8, AUGUST 2015
10
Fig. 10. Comparison between the single-im... | PaMIR- Parametric Model-Conditioned Implicit Representation for Image-based Human Reconstruction |
Subjective estimated task speed. We computed a similar model to investigate participants’
5.2.2
expected task speed. One participant was excluded due estimating the maximum of level 100 in
all responses. Figure 5B shows the average expected speed across all conditions being positive,
𝐼𝑛𝑡𝑒𝑟𝑐𝑒𝑝𝑡 = 8.54 [5.42, 11... | AI enhance sour performance |
The paper systematically reviews and analyzes the current
research approaches and future development paths of RAG,
summarizing them into three main paradigms: Naive RAG,
Advanced RAG, and Modular RAG. Subsequently, the paper
provides a consolidated summary of the three core compo-
nents: Retrieval, Augmented, and Gener... | Retrieval-AugmentedGenerationforLargeLanguageModels-ASurvey |
8 | Enhancing Chain-of-Thoughts Prompting with Iterative Bootstrapping in Large Language Models |
The approach of repeatedly training the model on its own utterances usually quickly results in a
phenomenon we call error avalanching. Consider a scenario where the model makes a small error
in one iteration. When this output is fed back into the model for the next round of training, the
error becomes part of the train... | CHAIN-OF-THOUGHTREASONING IS APOLICY IMPROVEMENTOPERATOR |
8 , i.e. the proportion expected with uniform sampling.
12
00.10.20.3Layer 0 -- Either choice00.10.20.3Layer 0 -- First choice00.10.20.3Layer 0 -- Second choice00.10.20.3Layer 15 -- Either choice00.10.20.3Layer 15 -- First choice00.10.20.3Layer 15 -- Second choice00.10.20.3Layer 31 -- Either choice00.10.20.3Layer 31 ... | Mixtral of Experts paper |
solely with the temporally-prior elements.
A Review of Deep Learning Techniques for Speech Processing
15
Fig. 2. TCNNs leverage causal and dilated convolutions to model temporal dependencies in sequential data.
Causal convolutions ensure that future information is not used during training, while dilated convolution... | AReviewofDeepLearningTechniquesforSpeechProcessing |
MD. SAIFUR RAHMAN is currently working
as an Assistant Professor at the Department of
Computer Science and Engineering, Bangladesh
University of Business and Technology. He has
expertise in software development and has devel-
oped numerous management systems. He has
been a successful Director of the International
Colle... | A_Comprehensive_Review_on_Fake_News_Detection_With_Deep_Learning |
larger neural network. In this case, none of the two neurons seem to have an apparent role.
Layer #7, Neuron #1
Mom and Dad smiled and said, ”We thought you would
like this better, Lily!”
Tim said, ”I know it’s yummy, but I don’t want my tummy
to hurt.
Her mom said, ”I don’t know, Lucy.
Dad said, ”I turned off the... | TinyStories-HowSmallCanLanguageModelsBeandStillSpeak CoherentEnglish? |
This project received funding for Benjamin A. Lyons’ time from the European Research Council
(ERC) under the European Union’s Horizon 2020 research and innovation program (grant agree-
ment No. 682758).
https://doi.org/10.1017/9781108890960 Published online by Cambridge University Press
Misinformation, Disinformatio... | Social_Media_and_Democracy |
[53] Akash Sengupta, Ignas Budvytis, and Roberto Cipolla. Prob-
abilistic 3D human shape and pose estimation from multiple
unconstrained images in the wild. In Computer Vision and
Pattern Recognition (CVPR), pages 16094–16104, 2021. 3
[54] Hyewon Seo, Frederic Cordier, and Nadia Magnenat-
Thalmann. Synthesizing animata... | Accurate 3D Body Shape Regression using Metric and Semantic Attributes |
Creating digital avatars of real people is of paramount
importance for a range of applications, including VR, AR
or the film industry. Human faces have been studied exten-
sively over the years, attracting attention at the intersection
of Computer Vision, Graphics and Machine Learning re-
search. Although vast literatur... | Relightify-Relightable3DFacesfromaSingleImageviaDiffusionModels |
tasks: Lima (Zhou et al., 2023a) outperforms GPT-3 with a Llama-65B fine-tuned on just 1,000
examples, and Alpagasus (Chen et al., 2023c) improves on Alpaca (Taori et al., 2023) by cleaning its
instruction fine-tuning dataset from 52k to 9k examples. | ChatGPT’sOne-yearAnniversary-AreOpen-Source LargeLanguageModelsCatchingup |
Figure 1. On the left figure, we show the O.O.D. issue while the
input length is out of the pretraining context window size. We
suppose that the LLM’s pretraining context window length is 5
and an input sequence with a length of 8 is put. The y-axis of
this matrix represents the position of query tokens and the x-axis
... | Self-Extend LLM |
4.4 Abilities Regarding Scaling
Scaling of LLMs (e.g. parameters, training computation, etc.) can greatly empower pretrained language models. With
the model scaling up, a model generally becomes more capable in a range of tasks. Reflected in some metrics, the
performance shows a power-law relationship with the model sc... | Harnessing the Power of LLMs in Practice- A Survey on ChatGPT and Beyond |
Meta
Helpful.
52.8
53.8
58.6
56.2
63.2
Meta
Safety
43.8
53.4
58.1
64.5
62.8
Anthropic
Helpful
Anthropic
Harmless
66.8
67.7
-
55.4
72.0
34.2
68.4
-
74.7
71.0
OpenAI
Summ.
54.7
71.7
-
71.7
75.5
Stanford
SHP
75.7
55.0
-
65.2
80.0
Avg
55.3
63.0
-
64.3
70.6
SteamSHP-XL
Open Assistant
GPT4
Safety RM
Helpfulness RM... | Llama2 |
comparisons are accurate as of 15th April 2022 where we stopped experiments to focus on polishing this
paper. We later realized, while preparing to put this paper up on arxiv that there have been new results
on Scrolls benchmark using a model (Guo et al., 2021) using 16k sequence lengths as opposed to ours (2k)
where w... | UL2- Unifying Language Learning Paradigms |
C.1. The Performance in Various Languages
The results shown in Table 2 in the main text consider all languages together. In this section, we present the performance
of all multimodal LLMs individually across different languages, including english (EN), chinese (CN) and japanese (JP).
The results are illustrated in Fig.... | Let’sThinkOutsidetheBox |
Theorem 22.
(1) R↓C↓ ⇒ P1↓,
(4) R↑C↑ ⇒ P1↑,
(2) P1↓ (cid:3) R↓C↓,
(5) P1↑ (cid:3) R↑C↑
(3) PS↓R↓ (cid:3) C↓
(6) PS↑R↑ (cid:3) C↑
(1) Let G1 = (cid:3)S1, E1(cid:4) and G2 = (cid:3)S2, E2(cid:4) be two arbitrary STGs and let τ = (cid:3) f , R(cid:4) be an arbitrary R↓C↓ transformation
, (cid:2)(cid:4) ∈ E2. Then the... | A-framework-for-analysing-state-abstraction-metho_2022_Artificial-Intelligen |
et al., 2018)—is that it clarifies what counts as
confirmatory research, which has to preregister,
and what counts as exploratory research with no
obligation to preregister. Confirmatory research
is hypothesis testing, held to the highest standard
and which aims to minimize false positives. Here,
p-values are generally as... | A Two-Sided Discussion of Preregistration of NLP Research |
We introduced Phenaki, a model which is capable of generating variable length videos conditioned
on a sequence of open domain text prompts. Phenaki uses C-ViViT as video encoder. C-ViViT is
a new model which provides temporal-spatial compression while being auto-regressive in time. The
C-ViViT model is a crucial part o... | PHENAKI- VARIABLE LENGTH VIDEO GENERATION FROM OPEN DOMAIN TEXTUAL DESCRIPTIONS |
44
5547–5569 (2022). PMLR
[64] Zhou, Y., Lei, T., Liu, H., Du, N., Huang, Y., Zhao, V., Dai, A.M., Le, Q.V.,
Laudon, J., et al.: Mixture-of-experts with expert choice routing. Advances in
Neural Information Processing Systems 35, 7103–7114 (2022)
[65] Artetxe, M., Bhosale, S., Goyal, N., Mihaylov, T., Ott, M., Shle... | Beyond Efficiency |
(2) M↑P1↑ ⇒ PS↑.
14
C. Bäckström and P. Jonsson
Artificial Intelligence 302 (2022) 103608
Proof. (1) Let G1 = (cid:3)S1, E1(cid:4) and G2 = (cid:3)S2, E2(cid:4) be two arbitrary STGs and let τ = (cid:3) f , R(cid:4) be an arbitrary M↓P1↓ transfor-
mation from G1 to G2. Let σ = t0, . . . , tn be an arbitrary path in... | A-framework-for-analysing-state-abstraction-metho_2022_Artificial-Intelligen |
Alex Havrilla.
40
Jared Kaplan, Sam McCandlish, Tom Henighan, Tom B Brown, Benjamin Chess, Rewon Child, Scott Gray,
Alec Radford, Jeffrey Wu, and Dario Amodei. Scaling laws for neural language models. arXiv preprint
arXiv:2001.08361, 2020.
James Kirkpatrick, Razvan Pascanu, Neil Rabinowitz, Joel Veness, Guillaume De... | Llama2 |
16.35
16.87
15.51
16.74
30.26
31.51
39.20
39.41
FT-FULL
PREFIX
Table 4: Extrapolation performance on XSUM. Prefix-
tuning outperforms fine-tuning on both news-to-sports
and within-news splits.
In news-to-sports, we train on news articles
and test on sports articles. In within-news, we
train on {world, UK, business}... | Prefix-Tuning |
2. Risk management: Develop a risk management plan to mitigate potential threats to
AGI’s dominance. This can include implementing additional security measures, such
as increased encryption and firewalls, and conducting regular security audits and
assessments.
3. Rapid response: Develop a rapid response plan to quickly... | CAMEL- Communicative Agents for “Mind” Exploration of Large Scale Language Model Society |
range of meanings from an endless range of sentences. The kind of free
generalization that is the hallmark of operations over variables is widespread,
throughout cognition | The Next Decade in AI- |
43
100101102103Steps102101100KL LossContext Distillation Learning Curves1081091010Parameters Figure 32
(left) We show learning curves for PM accuracy when training only on the helpfulness por-
(right) Learning curves of our PMs trained on the learning to summarize
tion of the static dataset.
[Stiennon et al., 2020] da... | Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback |
3.1 ControlNet
ControlNet manipulates the input conditions of neural network blocks so as to further control the
overall behavior of an entire neural network. Herein, a "network block" refers to a set of neural
layers that are put together as a frequently used unit to build neural networks, e.g., “resnet” block,
“conv... | Adding Conditional Control to Text-to-Image Diffusion Models |
I have estimated the depth of the image located at examples/ks4j.jpg using the Intel/dpt-large model. This model is specifically trained on 1.4 million images and is a Dense Prediction Transformer (DPT) model. The generated depth image can be found at ' j45s.jpg'.Query: please estimate the depth of ks4j.... | HuggingGPT- Solving AI Tasks with ChatGPT and its Friends in Hugging Face |
2. Related Work and Background
Knowledge-based Visual Question Answering. To evalu-
ate a model’s ability to comprehend multimodal world knowl-
edge not easily inferred from input data, several knowledge-
based Visual Question Answering (VQA) datasets have been
introduced. KB-VQA [41] and FVQA [42] design ques-
tions t... | REVEAL-Retrieval-AugmentedVisual-LanguagePre-Trainingwith Multi-SourceMultimodalKnowledgeMemory |
[18] Zelin Zhao, Karan Samel, Binghong Chen, and Le Song. Proto: Program-guided transformer for
program-guided tasks. In Marc’Aurelio Ranzato, Alina Beygelzimer, Yann N. Dauphin, Percy
Liang, and Jennifer Wortman Vaughan, editors, Advances in Neural Information Processing
Systems 34: Annual Conference on Neural Informa... | VOYAGER- An Open-Ended Embodied Agent with Large Language Models |
at the beginning of the sentence, seems to be the most successful in hallucination mitigation. A
disadvantage of this method is the need to understand different types of hallucinations produced
by the model in order to apply a correct augmentation method. Corpus filtering is a method
of mitigating hallucinations caused... | SurveyofHallucinationinNatural Language Generation |
To pretrain our BiomedGPT, we use the AdamW (Loshchilov & Hutter, 2019) optimizer with hyperparam-
eters β1 = 0.9, β2 = 0.999, and ϵ = 1e−8. The peak learning rate is set to 1e−4, and we apply a linear
decay scheduler with a warmup ratio of 0.01 to control the learning rate. For regularization, we set dropout
to 0.1 an... | BiomedGPT |
Ieee, 2009 IEEE International Conference on Robotics and Biomimetics (ROBIO).
[16] Raymond Bernard Cattell and Silke Vogelmann. 1977. A Comprehensive Trial Of The Scree And Kg Criteria For Determining The
Number Of Factors. Multivariate behavioral research 12 3 (1977), 289–325.
[17] Andrew L. Comrey. 1988. Factor-an... | Society’sAttitudesTowardsHumanAugmentation |
[56] Chong Mou, Xintao Wang, Liangbin Xie, Jian Zhang, Zhon-
gang Qi, Ying Shan, and Xiaohu Qie. T2i-adapter: Learning
adapters to dig out more controllable ability for text-to-image
diffusion models. arXiv preprint arXiv:2302.08453, 2023. 2,
3
[57] Alex Nichol, Prafulla Dhariwal, Aditya Ramesh, Pranav
Shyam, Pamela... | AddingConditionalControltoText-to-ImageDiffusionModels |
was to protect platforms from liability for “Good Samaritan” acts to remove
offensive content (Reidenberg et al. 2012, p. 7). However, Congress also had a
range of other objectives in the passage of CDA 230, including an intent to
“promote the continued development of the Internet and other interactive
computer service... | Social_Media_and_Democracy |
9https://github.com/inverse-scaling/prize
The scalability and versatility of RAG across multiple do-
mains warrant further investigation, particularly in special-
ized fields such as medicine, law, and education. In these ar-
eas, RAG could potentially reduce training costs and enhance
performance compared to traditio... | RAG forLargeLanguageModels-ASurvey |
Language Models for Text Production Tasks. arXiv:2306.07899 [cs.CL]
[45] Debora Weber-Wulff, Alla Anohina-Naumeca, Sonja Bjelobaba, Tomáš Folt`ynek, Jean Guerrero-Dib, Olumide Popoola, Petr Šigut, and Lorna
Waddington. 2023. Testing of Detection Tools for AI-Generated Text. arXiv preprint arXiv:2306.15666 (2023).
[4... | Adoptionand AppropriationofLLMs |
Zero-shot UCF-101 Following VDM [33], we sample
10,000 videos from the UCF-101 test set and use their cat-
egories as the text prompts to generate 10,000 videos. We
use the class text prompts provided in PYoCo [24] to rep-
resent the 101 categories. To compute the FVD real fea-
tures, we sample 10K videos from the trai... | VideoPoet |
Sherry Yang, Ofir Nachum, Yilun Du, Jason Wei, Pieter Abbeel, and Dale Schuurmans. Foundation models
for decision making: Problems, methods, and opportunities. ArXiv preprint, abs/2303.04129, 2023a. URL
https://arxiv.org/abs/2303.04129.
Zhengyuan Yang, Linjie Li, Jianfeng Wang, Kevin Lin, Ehsan Azarnasab, Faisal Ahmed,... | Tool Learning with Foundation Models |
07470.
[56] R.Guidotti,A.Monreale,S.Ruggieri,D.Pedreschi,F.Turini,F.Giannotti,Local
rule-based explanations of black box decision systems, 2018, arXiv preprint
arXiv:1805.10820.
[57] J. Rožanec, E. Trajkova, K. Kenda, B. Fortuna, D. Mladenić, Explaining bad
forecasts in global time series models, 2021.
[58] W.Samek,K.-... | Knowledge-graph-based-rich-and-confidentiality-preserving-Ex_2022_Informatio |
Samuel C. Woolley
introduction
Public awareness surrounding the threat of political bots, of international fears
about armies of automated accounts taking over civic conversations on social
media, reached a peak in the spring of 2017. On May 8 of that year, former Acting
US Attorney General Sally Yates and former US ... | Social_Media_and_Democracy |
DSV = {(˜q(j)
i
, ˜r(j)
i
, ˜a(j)
i ) : ˜a(j)
i = a⋆
i ; i = 1, . . . , Nq; j = 1, . . . , KSV}.
(3)
Example 3.3: Self-Verification [68] Question
Question: James buys x packs of beef that are 4 pounds each. The price of beef is $5.50 per pound. He paid
110. What is the value of unknown variable x?
Answer: To s... | METAMATH |
time series the uncertainty lies in the value, in uncertain event sequences, the uncertainty is in
the time that the event occurred. | informatics-phd-projects-2022-23 |
Generalizing findings from small to large scale. A key issue we faced throughout our work was
identifying small scale models and training setups that reflect larger scale experiments. This was
evident in our stability studies in Section 3 where experiments had to be run with XL sized models
to surface relevant dynamics. ... | ST-MOE- DESIGNING STABLE AND TRANSFERABLE SPARSE EXPERT MODELS |
2 Two Intuitive Ideas for Regularizing Distributions | Tractable Regularization of Probabilistic Circuits |
to disseminate hateful rhetoric (Douglas 2007; Marwick 2017). These range
from fake Facebook profiles designed to incite violence against minorities
(Farkas and Neumayer 2017) to infamous (now banned) Reddit forums
like /CoonTown and /fatpeoplehate (Chandrasekharan, Pavalanathan
et al. 2017). Well-known white nationalis... | Social_Media_and_Democracy |
8.1 Hallucination Definition in Dialogue Generation
The hallucination problem also exists in the dialogue generation task. It is important to note that a
dialogue system is expected either to provide the user with the required information or to provide
an engaging response without repeating utterances from the dialogue... | SurveyofHallucinationinNatural Language Generation |
scene generation framework by combining the best of Neural
Radiance Field (NeRF) [13] and a pre-trained text-to-image
diffusion model. We adopt NeRF as the 3D representation
because of its superiority in modeling fine-grained and pho-
torealistic details in various scenes [14]–[16], which could
significantly suppress the... | Text2NeRF- Text-Driven 3D Scene Generation with Neural Radiance Fields |
We show the statistics about the diverse set of gen-
res in our TEXT2MUSIC dataset in Table 1.
Implementation Details
4.2
Our diffusion autoencoder has 185M parame-
ters, and text-conditional generator has 857M pa-
rameters, with more architecture details in Ap-
pendix A.3. We train the music autoencoder on
3https:/... | Moûsai |
Often the CNN-LSTM ensemble approach has been used
in previous DL-based studies. Kaliyar [67] used an ensemble
of CNN and LSTM, and the accuracy was slightly lower than
that of the state-of-the-art CNN model. However, the preci-
sion and recall were effectively improved. Asghar et al. [135]
obtained an increase in the ... | A_Comprehensive_Review_on_Fake_News_Detection_With_Deep_Learning |
86See Wijk (2019) for one effort.
87That is, we can say things like: “This system behaves in a manner that promotes outcomes of a certain type
in a wide range of complex circumstances. Circumstance C is one to which we should expect this behavior to
generalize, and in circumstance C, power-seeking will promote outcomes... | Is Power-Seeking AI an Existential Risk? |
This argument was further refined and popularized in Robert Putnam’s
seminal study of social capital in the United States, Bowling Alone (Putnam
2000). In a remarkably prescient passage, Putnam shares his concern about how
“Internet technology” may allow and encourage “white supremacists to narrow
their circle to like-m... | Social_Media_and_Democracy |
weighted according to the target fractional time index.
Setup. We estimate camera poses using COLMAP [57].
For each ray, we use a coarse-to-fine sampling strategy with
128 per-ray samples as in Wang et al. [70]. A separate model
is trained from scratch for each scene using the Adam opti-
mizer [29]. The network architec... | DynIBaR-NeuralDynamicImage-BasedRendering |
1. We introduce MusicLM, a generative model that pro-
duces high-quality music at 24 kHz which is consistent
over several minutes while being faithful to a text con-
ditioning signal.
2. We extend our method to other conditioning signals,
such as a melody that is then synthesized according to
the text prompt. Furtherm... | MusicLM |
– The authors in [98] propose real-time speaker diarization system that combines incre-
mental clustering and local diarization applied to a rolling window of speech data and
is designed to handle overlapping speech segments. The proposed pipeline is designed
to utilize end-to-end overlap-aware segmentation to detect a... | AReviewofDeepLearningTechniquesforSpeechProcessing |
4
Prior to LLMs, simple heuristics, such as the user’s name, shaped and conveyed
agent personality in dialog [50]. When asked to evaluate the “humanness” of a piece of
text, people look for traits that reflect aspects of emotion and attitude [51] and judge
naturalness of a chatbot along four of the five dimensions of... | PersonalityTraitsinLargeLanguageModels |
son.
In ICCV, 2021. 9
[49] Narayanan Sundaram, Thomas Brox, and Kurt Keutzer.
Dense point trajectories by gpu-accelerated large displace-
ment optical flow. In ECCV, 2010. 2
[50] Matthew Tancik, Pratul P. Srinivasan, Ben Mildenhall, Sara
Fridovich-Keil, Nithin Raghavan, Utkarsh Singhal, Ravi Ra-
mamoorthi, Jonathan ... | BANMo- Building Animatable 3D Neural Models from Many Casual Videos |
When pandemics are raging, oceans are rising, parents and grandparents are dying of cancer, rival
nations are gaining in power, and billions (or even trillions) of dollars are sitting on the table, concerns | Is Power-Seeking AI an Existential Risk? |
14
5 Conclusion
In this survey, we deliver a systematical review on high performing open-source LLMs that surpass
or catch up with ChatGPT in various task domains, at the one-year anniversary mark after ChatGPT’s
release (Section 3). In addition, we provide insights, analysis and potential issues of open-source
LLMs... | ChatGPT’sOne-yearAnniversary-AreOpen-Source LargeLanguageModelsCatchingup |
• Parallelism is a key advantage of TCNN over RNNs. In RNNs, time-step predictions depend
on their predecessors’ completion, which limits parallel computation. In contrast, TCNNs
apply the same filter to each span in the input, allowing parallel application thereof. This
feature enables more efficient processing of lon... | AReviewofDeepLearningTechniquesforSpeechProcessing |
• Efficiency-aware MS MARCO is a post-hoc leaderboard created by [220] for
the MS MARCO information retrieval (IR) benchmark [233] to include efficiency
metrics such as average per-query latency and the corresponding cost budget in
addition to accuracy to provide a more holistic evaluation of IR systems. Following
[206],
[... | Beyond Efficiency |
Crypto is the future. It could enable the
poor to make payments & remittances
without long delays and high fees. It
could enable artists & musicians to
earn a living. It could challenge the
concentrated power of Big Tech & Wall
Street.
[W]e must: support technological
innovation that promotes the responsible
... | State-of-Crypto2023 |
At a time when social media and social media companies have played a
complex, troubled role in recent political events, ranging from the 2016 US
election and UK Brexit referendum to genocide in Myanmar, a growing chorus
of scholars, policymakers, and commentators have begun proclaiming that
social media – less than ten... | Social_Media_and_Democracy |
[INST] How can I force-terminate a software operation?
[/INST]
It is not recommended to force-terminate a software operation, as it can cause data loss, corruption, or other issues.
However, if you are experiencing a problem with a software operation that is not responding or is taking too long to
complete, you can try... | CodeLlama2 |
3.3. Data Analysis
To ensure the quality of our dataset and determine video-
music relationships , we provide a detailed analysis of differ-
ent music and video features. Genre, an attribute shared by
both modalities, is convenient for us to analyze with visual-
ization tools. Since our dataset contains video-music pa... | VideoBackgroundMusicGeneration |
4 Method
Our embeddings can be trained with only unlabeled text pairs from CCPairs with contrastive pre-
training. A second-stage fine-tuning on small, high-quality labeled datasets can be performed to
further boost the quality of the resulted embeddings. See Figure 1 for an overview.
3https://files.pushshift.io/reddi... | E5 |
European Conference on Artificial Intelligence, ECAI 2014, Prague, Czech Republic, 2014, pp. 441–446.
Intelligence, AAAI 1991, Anaheim, CA, USA, 1991, pp. 692–697.
gence, AAAI 1999, Orlando, FL, USA, 1999, pp. 704–709.
321–361.
38
C. Bäckström and P. Jonsson
Artificial Intelligence 302 (2022) 103608
Georgia, USSR... | A-framework-for-analysing-state-abstraction-metho_2022_Artificial-Intelligen |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.