Unnamed: 0
int64 2
9.3k
| sentence
stringlengths 30
941
| aspect_term_1
stringlengths 1
32
⌀ | aspect_term_2
stringlengths 2
27
⌀ | aspect_term_3
stringlengths 2
23
⌀ | aspect_term_4
stringclasses 25
values | aspect_term_5
stringclasses 7
values | aspect_term_6
stringclasses 1
value | aspect_category_1
stringclasses 9
values | aspect_category_2
stringclasses 9
values | aspect_category_3
stringclasses 9
values | aspect_category_4
stringclasses 2
values | aspect_category_5
stringclasses 1
value | aspect_term_1_polarity
stringclasses 3
values | aspect_term_2_polarity
stringclasses 3
values | aspect_term_3_polarity
stringclasses 3
values | aspect_term_4_polarity
stringclasses 3
values | aspect_term_5_polarity
stringclasses 3
values | aspect_term_6_polarity
stringclasses 1
value | aspect_category_1_polarity
stringclasses 3
values | aspect_category_2_polarity
stringclasses 3
values | aspect_category_3_polarity
stringclasses 3
values | aspect_category_4_polarity
stringclasses 1
value | aspect_category_5_polarity
stringclasses 1
value |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
9,115 | In particular, the advantages and disadvantages of different categories are not systematically compared, and hence the readers cannot get insightful comments and suggestions from this survey.[advantages and disadvantages-NEG], [EMP-NEG] | advantages and disadvantages | null | null | null | null | null | EMP | null | null | null | null | NEG | null | null | null | null | null | NEG | null | null | null | null |
9,116 | n In general, survey papers are not very suitable for publication at conferences.[survey papers-NEG, conferences-NEG], [APR-NEG]] | survey papers | conferences | null | null | null | null | APR | null | null | null | null | NEG | NEG | null | null | null | null | NEG | null | null | null | null |
9,118 | It makes several important contributions, including extending the previously published bounds by Telgarsky et al. to tighter bounds for the special case of ReLU DNNs, giving a construction for a family of hard functions whose affine pieces scale exponentially with the dimensionality of the inputs, and giving a procedure for searching for globally optimal solution of a 1-hidden layer ReLU DNN with linear output layer and convex loss.[contributions-POS], [EMP-POS] | contributions | null | null | null | null | null | EMP | null | null | null | null | POS | null | null | null | null | null | POS | null | null | null | null |
9,119 | I think these contributions warrant publishing the paper at ICLR 2018.[contributions-POS, paper-POS], [APR-POS, REC-POS] | contributions | paper | null | null | null | null | APR | REC | null | null | null | POS | POS | null | null | null | null | POS | POS | null | null | null |
9,120 | The paper is also well written, a bit dense in places, but overall well organized and easy to follow.[paper-POS], [CLA-POS, PNF-POS] | paper | null | null | null | null | null | CLA | PNF | null | null | null | POS | null | null | null | null | null | POS | POS | null | null | null |
9,121 | A key limitation of the paper in my opinion is that typically DNNs do not contain a linear final layer.[limitation-NEG, paper-NEG], [EMP-NEG] | limitation | paper | null | null | null | null | EMP | null | null | null | null | NEG | NEG | null | null | null | null | NEG | null | null | null | null |
9,122 | It will be valuable to note what, if any, of the representation analysis and global convergence results carry over to networks with non-linear (Softmax, e.g.) final layer.[representation analysis-NEU, results-NEU], [EMP-NEU] | representation analysis | results | null | null | null | null | EMP | null | null | null | null | NEU | NEU | null | null | null | null | NEU | null | null | null | null |
9,123 | I also think that the global convergence algorithm is practically unfeasible for all but trivial use cases due to terms like D^nw, would like hearing authors' comments in case I'm missing some simplification.[algorithm-NEG], [EMP-NEG] | algorithm | null | null | null | null | null | EMP | null | null | null | null | NEG | null | null | null | null | null | NEG | null | null | null | null |
9,124 | One minor suggestion for improving readability is to explicitly state, whenever applicable, that functions under consideration are PWL.[null], [SUB-NEU, EMP-NEU] | null | null | null | null | null | null | SUB | EMP | null | null | null | null | null | null | null | null | null | NEU | NEU | null | null | null |
9,125 | For example, adding PWL to Theorems and Corollaries in Section 3.1 will help. [Theorems-NEU, Section-NEU], [EMP-NEU] | Theorems | Section | null | null | null | null | EMP | null | null | null | null | NEU | NEU | null | null | null | null | NEU | null | null | null | null |
9,126 | Similarly would be good to state, wherever applicable, the DNN being discussed is a ReLU DNN.[null], [SUB-NEU] | null | null | null | null | null | null | SUB | null | null | null | null | null | null | null | null | null | null | NEU | null | null | null | null |
9,130 | I have two problems with these claims: 1) Modern ConvNet architectures (Inception, ResNeXt, SqueezeNet, BottleNeck-DenseNets and ShuffleNets) don't have large fully connected layers.[claims-NEG], [EMP-NEG] | claims | null | null | null | null | null | EMP | null | null | null | null | NEG | null | null | null | null | null | NEG | null | null | null | null |
9,131 | 2) The authors reject the technique of 'Deep compression' as being impractical.[technique-NEU], [EMP-NEU] | technique | null | null | null | null | null | EMP | null | null | null | null | NEU | null | null | null | null | null | NEU | null | null | null | null |
9,132 | I suspect it is actually much easier to use in practice as you don't have to a-priori know the correct level of sparsity for every level of the network.[null], [EMP-NEG] | null | null | null | null | null | null | EMP | null | null | null | null | null | null | null | null | null | null | NEG | null | null | null | null |
9,133 | p3. What does 'normalized' mean?[null], [EMP-NEU] | null | null | null | null | null | null | EMP | null | null | null | null | null | null | null | null | null | null | NEU | null | null | null | null |
9,135 | p3. Are you using an L2 weight penalty?[null], [EMP-NEU] | null | null | null | null | null | null | EMP | null | null | null | null | null | null | null | null | null | null | NEU | null | null | null | null |
9,136 | If not, your fully-connected baseline may be unnecessarily overfitting the training data.[baseline-NEG], [EMP-NEG] | baseline | null | null | null | null | null | EMP | null | null | null | null | NEG | null | null | null | null | null | NEG | null | null | null | null |
9,137 | p3. Table 1. Where do the choice of CL Junction densities come from?[Table-NEU], [EMP-NEU] | Table | null | null | null | null | null | EMP | null | null | null | null | NEU | null | null | null | null | null | NEU | null | null | null | null |
9,138 | Did you do a grid search to find the optimal level of sparsity at each level?[null], [EMP-NEU] | null | null | null | null | null | null | EMP | null | null | null | null | null | null | null | null | null | null | NEU | null | null | null | null |
9,139 | p7-8. I had trouble following the left/right & front/back notation.[p-NEU], [PNF-NEG] | p | null | null | null | null | null | PNF | null | null | null | null | NEU | null | null | null | null | null | NEG | null | null | null | null |
9,140 | p8. Figure 7. How did you decide which data points to include in the plots?[p-NEU, Figure-NEU], [EMP-NEU] | p | Figure | null | null | null | null | EMP | null | null | null | null | NEU | NEU | null | null | null | null | NEU | null | null | null | null |
9,142 | Congratulations on a very interesting and clear paper.[paper-POS], [CLA-POS] | paper | null | null | null | null | null | CLA | null | null | null | null | POS | null | null | null | null | null | POS | null | null | null | null |
9,143 | While ICLR is not focused on neuroscientific studies, this paper clearly belongs here as it shows what representations develop in recurrent networks that are trained on spatial navigation.[paper-POS], [APR-POS] | paper | null | null | null | null | null | APR | null | null | null | null | POS | null | null | null | null | null | POS | null | null | null | null |
9,145 | I found it is very interesting that the emergence of these representations was contingent on some regularization constraint.[representations-POS], [EMP-POS] | representations | null | null | null | null | null | EMP | null | null | null | null | POS | null | null | null | null | null | POS | null | null | null | null |
9,146 | This seems similar to the visual domain where edge detectors emerge easily when trained on natural images with sparseness constraints as in Olshausen&Field and later reproduced with many other models that incorporate sparseness constraints.[null], [EMP-POS] | null | null | null | null | null | null | EMP | null | null | null | null | null | null | null | null | null | null | POS | null | null | null | null |
9,147 | I do have some questions about the training itself.[training-NEU], [EMP-NEU] | training | null | null | null | null | null | EMP | null | null | null | null | NEU | null | null | null | null | null | NEU | null | null | null | null |
9,148 | The paper mentions a metabolic cost that is not specified in the paper.[paper-NEG], [SUB-NEG] | paper | null | null | null | null | null | SUB | null | null | null | null | NEG | null | null | null | null | null | NEG | null | null | null | null |
9,149 | This should be added.[null], [SUB-NEG] | null | null | null | null | null | null | SUB | null | null | null | null | null | null | null | null | null | null | NEG | null | null | null | null |
9,151 | I am puzzled why is the error is coming down before the boundary interaction?[error-NEU], [EMP-NEU] | error | null | null | null | null | null | EMP | null | null | null | null | NEU | null | null | null | null | null | NEU | null | null | null | null |
9,152 | Even more puzzling, why does this error go up again for the blue curve (no interaction)? Shouldn't at least this curve be smooth? [error-NEU], [EMP-NEU] | error | null | null | null | null | null | EMP | null | null | null | null | NEU | null | null | null | null | null | NEU | null | null | null | null |
9,155 | On the positive side, the paper is mostly well-written, seems technically correct, and there are some results that indicate that the MSA is working quite well on relatively complex tasks.[paper-POS, results-POS], [CLA-POS, EMP-POS] | paper | results | null | null | null | null | CLA | EMP | null | null | null | POS | POS | null | null | null | null | POS | POS | null | null | null |
9,156 | On the negative side, there seems to be relatively limited novelty: we can think of MSA as one particular communication (i.e, star) configuration one could use is a multiagent system.[novelty-NEG], [NOV-NEG] | novelty | null | null | null | null | null | NOV | null | null | null | null | NEG | null | null | null | null | null | NEG | null | null | null | null |
9,157 | One aspect does does strike me as novel is the gated composition module, which allows differentiation of messages to other agents based on the receivers internal state.[gated composition module-POS], [NOV-POS] | gated composition module | null | null | null | null | null | NOV | null | null | null | null | POS | null | null | null | null | null | POS | null | null | null | null |
9,158 | (So, the *interpretation* of the message is learned). I like this idea,[idea-POS], [EMP-POS] | idea | null | null | null | null | null | EMP | null | null | null | null | POS | null | null | null | null | null | POS | null | null | null | null |
9,159 | however, the results are mixed, and the explanation given is plausible, but far from a clearly demonstrated answer.[results-NEU, explanation-NEU], [EMP-NEG] | results | explanation | null | null | null | null | EMP | null | null | null | null | NEU | NEU | null | null | null | null | NEG | null | null | null | null |
9,161 | however the summed global signal is hand crafted information and does not facilitate an independently reasoning master agent.[issues-NEU], [SUB-NEU] | issues | null | null | null | null | null | SUB | null | null | null | null | NEU | null | null | null | null | null | NEU | null | null | null | null |
9,162 | -Please explain what is meant here by 'hand crafted information', my understanding is that the f^i in figure 1 of that paper are learned modules?[figure-NEU, modules-NEU], [PNF-NEU] | figure | modules | null | null | null | null | PNF | null | null | null | null | NEU | NEU | null | null | null | null | NEU | null | null | null | null |
9,163 | -Please explain what would be the differences with CommNet with 1 extra agent that takes in the same information as your 'master'.[null], [EMP-NEU] | null | null | null | null | null | null | EMP | null | null | null | null | null | null | null | null | null | null | NEU | null | null | null | null |
9,164 | *This relates also to this: Later we empirically verify that, even when the overall in- formation revealed does not increase per se, an independent master agent tend to absorb the same information within a big picture and effectively helps to make decision in a global manner.[null], [EMP-NEU] | null | null | null | null | null | null | EMP | null | null | null | null | null | null | null | null | null | null | NEU | null | null | null | null |
9,167 | Specifically, we compare the performance among the CommNet model, our MS-MARL model without explicit master state (e.g. the occupancy map of controlled agents in this case), and our full model with an explicit occupancy map as a state to the master agent.[performance-NEU], [EMP-NEU] | performance | null | null | null | null | null | EMP | null | null | null | null | NEU | null | null | null | null | null | NEU | null | null | null | null |
9,168 | As shown in Figure 7 (a)(b), by only allowed an independently thinking master agent and communication among agents, our model already outperforms the plain CommNet model which only supports broadcast- ing communication of the sum of the signals.[model-POS], [EMP-NEU] | model | null | null | null | null | null | EMP | null | null | null | null | POS | null | null | null | null | null | NEU | null | null | null | null |
9,169 | -Minor: I think that the statement which only supports broadcast-ing communication of the sum of the signals is not quite fair: surely they have used a 1-channel communication structure, but it would be easy to generalize that.[statement-NEG], [EMP-NEG] | statement | null | null | null | null | null | EMP | null | null | null | null | NEG | null | null | null | null | null | NEG | null | null | null | null |
9,170 | -Major: When I look at figure 4D, I see that the proposed approach *also* only provides the master with the sum (or really mean) with of the individual messages...? So it is not quite clear to me what explains the difference. *In 4.4, it is not quite clear exactly how the figure of master and slave actions is created.[proposed approach-NEG, figure-NEU], [EMP-NEU] | proposed approach | figure | null | null | null | null | EMP | null | null | null | null | NEG | NEU | null | null | null | null | NEU | null | null | null | null |
9,171 | This seems to suggest that the only thing that the master can communicate is action information?[null], [EMP-NEU] | null | null | null | null | null | null | EMP | null | null | null | null | null | null | null | null | null | null | NEU | null | null | null | null |
9,173 | * In table 2, it is not clear how significant these differences are.[table-NEG], [PNF-NEG] | table | null | null | null | null | null | PNF | null | null | null | null | NEG | null | null | null | null | null | NEG | null | null | null | null |
9,174 | What are the standard errors?[standard errors-NEU], [EMP-NEG] | standard errors | null | null | null | null | null | EMP | null | null | null | null | NEU | null | null | null | null | null | NEG | null | null | null | null |
9,175 | * The section 3.2 explains standard things (policy gradient), but the details are a bit unclear.[section-NEG], [SUB-NEG] | section | null | null | null | null | null | SUB | null | null | null | null | NEG | null | null | null | null | null | NEG | null | null | null | null |
9,176 | In particular, I do not see how the Gaussian/softmax layers are integrated; they do not seem to appear in figure 4?[figure-NEG], [SUB-NEG] | figure | null | null | null | null | null | SUB | null | null | null | null | NEG | null | null | null | null | null | NEG | null | null | null | null |
9,177 | * I cannot understand figure 7 without more explanation.[figure-NEG, explanation-NEG], [SUB-NEG] | figure | explanation | null | null | null | null | SUB | null | null | null | null | NEG | NEG | null | null | null | null | NEG | null | null | null | null |
9,178 | (The background is all black - did something go wrong with the pdf?)[background-NEG], [PNF-NEG] | background | null | null | null | null | null | PNF | null | null | null | null | NEG | null | null | null | null | null | NEG | null | null | null | null |
9,179 | Details: * references are wrongly formatted throughout.[references-NEG], [PNF-NEG] | references | null | null | null | null | null | PNF | null | null | null | null | NEG | null | null | null | null | null | NEG | null | null | null | null |
9,180 | * In this regard, we are among the first to combine both the centralized perspective and the decentralized perspective This is a weak statement (E.g., I suppose that in the greater scheme of things all of us will be amongst the first people that have walked this earth...)[null], [NOV-NEG] | null | null | null | null | null | null | NOV | null | null | null | null | null | null | null | null | null | null | NEG | null | null | null | null |
9,183 | Can it be made crisper?[null], [PNF-NEU] | null | null | null | null | null | null | PNF | null | null | null | null | null | null | null | null | null | null | NEU | null | null | null | null |
9,184 | * Note here that, although we explicitly input an occupancy map to the master agent, the actual infor- mation of the whole system remains the same.[information-NEU], [EMP-NEU] | information | null | null | null | null | null | EMP | null | null | null | null | NEU | null | null | null | null | null | NEU | null | null | null | null |
9,185 | This is a somewhat peculiar statement.[statement-NEG], [PNF-NEG] | statement | null | null | null | null | null | PNF | null | null | null | null | NEG | null | null | null | null | null | NEG | null | null | null | null |
9,186 | Clearly, the distribution of information over the agents is crucial.[information-NEU], [EMP-NEU] | information | null | null | null | null | null | EMP | null | null | null | null | NEU | null | null | null | null | null | NEU | null | null | null | null |
9,191 | This works because each variable (of the state space) is modified in turn, so that the resulting update is invertible, with a tractable transformation inspired by Dinh et al 2016.[variable-NEU, update-NEU], [CMP-NEU] | variable | update | null | null | null | null | CMP | null | null | null | null | NEU | NEU | null | null | null | null | NEU | null | null | null | null |
9,192 | Overall, I believe this paper is of good quality, clearly and carefully written, and potentially accelerates mixing in a state-of-the-art MCMC method, HMC, in many practical cases.[paper-POS], [CLA-POS, EMP-POS] | paper | null | null | null | null | null | CLA | EMP | null | null | null | POS | null | null | null | null | null | POS | POS | null | null | null |
9,194 | The experimental section proves the usefulness of the method on a range of relevant test cases; in addition, an application to a latent variable model is provided sec5.2.[section-POS, method-POS, sec-POS], [EMP-POS] | section | method | sec | null | null | null | EMP | null | null | null | null | POS | POS | POS | null | null | null | POS | null | null | null | null |
9,195 | Fig 1a presents results in terms of numbers of gradient evaluations, but I couldn't find much in the way of computational cost of L2HMC in the paper. [Fig-NEG, results-NEG, paper-NEU], [SUB-NEG, EMP-NEG] | Fig | results | paper | null | null | null | SUB | EMP | null | null | null | NEG | NEG | NEU | null | null | null | NEG | NEG | null | null | null |
9,196 | I can't see where the number 124x in sec 5.1 stems from.[sec-NEG], [CLA-NEG] | sec | null | null | null | null | null | CLA | null | null | null | null | NEG | null | null | null | null | null | NEG | null | null | null | null |
9,197 | As a user, I would be interested in the typical computational cost of both MCMC sampler training and MCMC sampler usage (inference?), compared to competing methods.[competing methods-NEU], [SUB-NEU] | competing methods | null | null | null | null | null | SUB | null | null | null | null | NEU | null | null | null | null | null | NEU | null | null | null | null |
9,198 | This is admittedly hard to quantify objectively, but just an order of magnitude would be helpful for orientation.[null], [SUB-NEU] | null | null | null | null | null | null | SUB | null | null | null | null | null | null | null | null | null | null | NEU | null | null | null | null |
9,199 | Would it be relevant, in sec5.1, to compare to other methods than just HMC, eg LAHMC?[sec-NEG], [CMP-NEG, SUB-NEG] | sec | null | null | null | null | null | CMP | SUB | null | null | null | NEG | null | null | null | null | null | NEG | NEG | null | null | null |
9,200 | I am missing an intuition for several things: eq7, the time encoding defined in Appendix C Appendix Fig5, I cannot quite see how the caption claim is supported by the figure (just hardly for VAE, but not for HMC).[eq-NEG, Appendix-NEG, Fig-NEG, figure-NEG], [PNF-NEG, CLA-NEG] | eq | Appendix | Fig | figure | null | null | PNF | CLA | null | null | null | NEG | NEG | NEG | NEG | null | null | NEG | NEG | null | null | null |
9,202 | # Minor errors - sec1: The sampler is trained to minimize a variation: should be maximize as well as on a the real-world[sec-NEG], [EMP-NEG] | sec | null | null | null | null | null | EMP | null | null | null | null | NEG | null | null | null | null | null | NEG | null | null | null | null |
9,203 | - sec3.2 and 1/2 v^T v the kinetic: energy missing[sec-NEG], [SUB-NEG] | sec | null | null | null | null | null | SUB | null | null | null | null | NEG | null | null | null | null | null | NEG | null | null | null | null |
9,204 | - sec4: the acronym L2HMC is not expanded anywhere in the paper[sec-NEG], [CLA-NEG, PNF-NEG] | sec | null | null | null | null | null | CLA | PNF | null | null | null | NEG | null | null | null | null | null | NEG | NEG | null | null | null |
9,205 | The sentence We will denote the complete augmented...p(d) might be moved to after from a uniform distribution in the same paragraph.[sentence-NEU, paragraph-NEU], [PNF-NEU] | sentence | paragraph | null | null | null | null | PNF | null | null | null | null | NEU | NEU | null | null | null | null | NEU | null | null | null | null |
9,206 | In paragraph starting We now update x: - specify for clarity: the first update, which yields x' / the second update, which yields x'' [paragraph-NEG], [CLA-NEG, PNF-NEG] | paragraph | null | null | null | null | null | CLA | PNF | null | null | null | NEG | null | null | null | null | null | NEG | NEG | null | null | null |
9,207 | - only affects $x_{bar{m}^t}$: should be $x'_{bar{m}^t}$ (prime missing) [null], [PNF-NEG] | null | null | null | null | null | null | PNF | null | null | null | null | null | null | null | null | null | null | NEG | null | null | null | null |
9,208 | - the syntax using subscript m^t is confusing to read; wouldn't it be clearer to write this as a function, eg mask(x',m^t)?[syntax-NEG], [PNF-NEG, CLA-NEG] | syntax | null | null | null | null | null | PNF | CLA | null | null | null | NEG | null | null | null | null | null | NEG | NEG | null | null | null |
9,209 | - inside zeta_2 and zeta_3, do you not mean $m^t and $bar{m}^t$ ?[null], [PNF-NEG] | null | null | null | null | null | null | PNF | null | null | null | null | null | null | null | null | null | null | NEG | null | null | null | null |
9,210 | - sec5: add reference for first mention of A NICE MC[sec-NEG, reference-NEG], [PNF-NEG] | sec | reference | null | null | null | null | PNF | null | null | null | null | NEG | NEG | null | null | null | null | NEG | null | null | null | null |
9,211 | - Appendix A: - Let's -> Let [Appendix-NEG], [PNF-NEG] | Appendix | null | null | null | null | null | PNF | null | null | null | null | NEG | null | null | null | null | null | NEG | null | null | null | null |
9,212 | - eq12 should be x'' ... -[eq-NEG], [PNF-NEG] | eq | null | null | null | null | null | PNF | null | null | null | null | NEG | null | null | null | null | null | NEG | null | null | null | null |
9,213 | Appendix C: space missing after Section 5.1[Appendix-NEG, Section-NEG], [PNF-NEG] | Appendix | Section | null | null | null | null | PNF | null | null | null | null | NEG | NEG | null | null | null | null | NEG | null | null | null | null |
9,214 | - Appendix D1: In this section is presented : sounds odd[Appendix-NEG, section-NEG], [PNF-NEG] | Appendix | section | null | null | null | null | PNF | null | null | null | null | NEG | NEG | null | null | null | null | NEG | null | null | null | null |
9,215 | n- Appendix D3: presumably this should consist of the figure 5 ? Maybe specify.[Appendix-NEG, figure-NEG], [PNF-NEG]] | Appendix | figure | null | null | null | null | PNF | null | null | null | null | NEG | NEG | null | null | null | null | NEG | null | null | null | null |
9,218 | Strengths: The proposed method has achieved a better convergence rate in different tasks than all other hand-engineered algorithms.[proposed method-POS], [EMP-POS] | proposed method | null | null | null | null | null | EMP | null | null | null | null | POS | null | null | null | null | null | POS | null | null | null | null |
9,219 | The proposed method has better robustess in different tasks and different batch size setting.[proposed method-POS], [EMP-POS] | proposed method | null | null | null | null | null | EMP | null | null | null | null | POS | null | null | null | null | null | POS | null | null | null | null |
9,220 | The invariant of coordinate permutation and the use of block-diagonal structure improve the efficiency of LQG.[null], [EMP-POS] | null | null | null | null | null | null | EMP | null | null | null | null | null | null | null | null | null | null | POS | null | null | null | null |
9,221 | Weaknesses: 1. Since the batch size is small in each experiment, it is hard to compare convergence rate within one epoch.[null], [EMP-NEU] | null | null | null | null | null | null | EMP | null | null | null | null | null | null | null | null | null | null | NEU | null | null | null | null |
9,222 | More iterations should be taken and the log-scale style figure is suggested.[null], [EMP-NEU] | null | null | null | null | null | null | EMP | null | null | null | null | null | null | null | null | null | null | NEU | null | null | null | null |
9,223 | 2. In Figure 1b, L2LBGDBGD converges to a lower objective value, while the other figures are difficult to compare, the convergence value should be reported in all experiments.[Figure-NEU, experiments-NEG], [CMP-NEG] | Figure | experiments | null | null | null | null | CMP | null | null | null | null | NEU | NEG | null | null | null | null | NEG | null | null | null | null |
9,224 | 3. "The average recent iterate" described in section 3.6 uses recent 3 iterations to compute the average, the reason to choose "3", and the effectiveness of different choices should be discussed, as well as the "24" used in state features.[section-NEU], [EMP-NEU] | section | null | null | null | null | null | EMP | null | null | null | null | NEU | null | null | null | null | null | NEU | null | null | null | null |
9,225 | 4. Since the block-diagonal structure imposed on A_t, B_t, and F_t, how to choose a proper block size?[null], [EMP-NEU] | null | null | null | null | null | null | EMP | null | null | null | null | null | null | null | null | null | null | NEU | null | null | null | null |
9,226 | Or how to figure out a coordinate group?[null], [EMP-NEU] | null | null | null | null | null | null | EMP | null | null | null | null | null | null | null | null | null | null | NEU | null | null | null | null |
9,227 | 5. The caption in Figure 1,3, "with 48 input and hidden units" should clarify clearly.[Figure-NEG], [CLA-NEG] | Figure | null | null | null | null | null | CLA | null | null | null | null | NEG | null | null | null | null | null | NEG | null | null | null | null |
9,228 | The curves of different methods are suggested to use different lines (e.g., dashed lines) to denote different algorithms rather than colors only.[null], [EMP-NEU] | null | null | null | null | null | null | EMP | null | null | null | null | null | null | null | null | null | null | NEU | null | null | null | null |
9,229 | 6. typo: sec 1 parg 5, "current iterate" -> "current iteration".[typo-NEG], [CLA-NEG] | typo | null | null | null | null | null | CLA | null | null | null | null | NEG | null | null | null | null | null | NEG | null | null | null | null |
9,231 | by Li & Malik, this paper tends to solve the high-dimensional problem.[paper-NEU], [CMP-NEU] | paper | null | null | null | null | null | CMP | null | null | null | null | NEU | null | null | null | null | null | NEU | null | null | null | null |
9,232 | With the new observation of invariant in coordinates permutation in neural networks, this paper imposes the block-diagonal structure in the model to reduce the complexity of LQG algorithm.[paper-NEU, model-NEU], [EMP-NEU] | paper | model | null | null | null | null | EMP | null | null | null | null | NEU | NEU | null | null | null | null | NEU | null | null | null | null |
9,239 | I could not find any technical contribution or something sufficiently mature and interesting for presenting in ICLR.[technical contribution-NEG], [APR-NEG] | technical contribution | null | null | null | null | null | APR | null | null | null | null | NEG | null | null | null | null | null | NEG | null | null | null | null |
9,240 | Some issues: - submission is supposed to be double blind but authors reveal their identity at the start of section 2.1.[section-NEG], [PNF-NEG] | section | null | null | null | null | null | PNF | null | null | null | null | NEG | null | null | null | null | null | NEG | null | null | null | null |
9,241 | - implementation details all over the place (section 3. is called Implementation, but at that point no concrete idea has been proposed, so it seems too early for talking about tensorflow and keras).[implementation details-NEG, section-NEG], [PNF-NEG, EMP-NEG] | implementation details | section | null | null | null | null | PNF | EMP | null | null | null | NEG | NEG | null | null | null | null | NEG | NEG | null | null | null |
9,245 | 2) though the non-saturating variant (see Eq. 3) of ``standard GAN'' may converge towards a minimum of the Jensen-Shannon divergence, it does not mean that the minimization process follows gradients of the Jensen-Shannon divergence (and conversely, following gradient paths of the Jensen-Shannon divergence may not converge towards a minimum, but this was rather the point of the previous critiques about ``standard GAN''). [null], [EMP-NEU] | null | null | null | null | null | null | EMP | null | null | null | null | null | null | null | null | null | null | NEU | null | null | null | null |
9,246 | 3) the penalization strategies introduced for ``non-standard GAN'' with specific motivations, may also apply successfully to the ``standard GAN'', improving robustness, thereby helping to set hyperparameters.[null], [EMP-NEU] | null | null | null | null | null | null | EMP | null | null | null | null | null | null | null | null | null | null | NEU | null | null | null | null |
9,248 | Overall, I believe that the paper provides enough material to substantiate these claims, even if the message could be better delivered.[claims-NEU], [SUB-POS] | claims | null | null | null | null | null | SUB | null | null | null | null | NEU | null | null | null | null | null | POS | null | null | null | null |
9,249 | In particular, the writing is sometimes ambiguous (e.g. in Section 2.3, the reader who did not follow the recent developments on the subject on arXiv will have difficulties to rebuild the cross-references between authors, acronyms and formulae).[writing-NEG], [CLA-NEG] | writing | null | null | null | null | null | CLA | null | null | null | null | NEG | null | null | null | null | null | NEG | null | null | null | null |