In the years before this paper came out in 2017, a number of different graph convolution architectures - which use weight-sharing and order-invariant operations to create representations at nodes in a graph that are contextualized by information in the rest of the graph - had been suggested for learning representations of molecules. The authors of this paper out of Google sought to pull all of these proposed models into a single conceptual framework, for the sake of better comparing and testing the design choices that went into them. All empirical tests were done using the QM9 dataset, where 134,000 molecules have predicted chemical properties attached to them, things like the amount of energy released if bombs are sundered and the energy of electrons at different electron shells. 

https://i.imgur.com/Mmp8KO6.png

An interesting note is that these properties weren't measured empirically, but were simulated by a very expensive quantum simulation, because the former wouldn't be feasible for this large of a dataset. However, this is still a moderately interesting test because, even if we already have the capability to computationally predict these features, a neural network would do much more quickly. And, also, one might aspirationally hope that architectures which learn good representations of molecules for quantum predictions are also useful for tasks with a less available automated prediction mechanism.

The framework assumes the existence of "hidden" feature vectors h at each node (atom) in the graph, as well as features that characterize the edges between nodes (whether that characterization comes through sorting into discrete bond categories or through a continuous representation). The features associated with each atom at the lowest input level of the molecule-summarizing networks trained here include: the element ID, the atomic number, whether it accepts electrons or donates them, whether it's in an aromatic system, and which shells its electrons are in.  

https://i.imgur.com/J7s0q2e.png

Given these building blocks, the taxonomy lays out three broad categories of function, each of which different architectures implement in slightly different ways. 

1. The Message function, M(). This function is defined with reference to a node w, that the message is coming from, and a node v, that it's being sent to, and is meant to summarize the information coming from w to inform the node representation that will be calculated at v. It takes into account the feature vectors of one or both nodes at the next level down, and sometimes also incorporates feature vectors attached to the  edge connecting the two nodes. In a notable example of weight sharing, you'd use the same Message function for every combination of v and w, because you need to be able to process an arbitrary number of pairs, with each v having a different number of neighbors.  The simplest example you might imagine here is a simple concatenation of incoming node and edge features; a more typical example from the architectures reviewed is a concatenation followed by a neural network layer. The aggregate message being sent to the receiver node is calculated by summing together the messages from each incoming vector (though it seems like other options are possible; I'm a bit confused why the paper presented summing as the only order-invariant option). 
2. The Update function, U(). This function governs how to take the aggregated message vector sent to a particular node, and combine that with the prior-layer representation at that node, to come up with a next-layer representation at that node. Similarly, the same Update function weights are shared across all atoms. 
3. The Readout function, R(), which takes the final-layer representation of each atom node and aggregates the representations into a final graph-level representation an order-invariant way 

Rather than following in the footsteps of the paper by describing each proposed model type and how it can be described in this framework, I'll instead try to highlight some of the more interesting ways in which design choices differed across previously proposed architectures. 

- Does the message function being sent from w to v depend on the feature value at both w and v, or just v? To put the question more colloquially, you might imagine w wanting to contextually send different information based on different values of the feature vector at node v, and this extra degree of expressivity (not present in the earliest 2015 paper), seems like a quite valuable addition (in that all subsequent papers include it)
- Are the edge features static, categorical things, or are they feature vectors that get iteratively updated in the same way that the node vectors do? For most of the architectures reviewed, the former is true, but the authors found that the highest performance in their tests came from networks with continuous edge vectors, rather than just having different weights for different category types of edge
- Is the Readout function something as simple as a summation of all top-level feature vectors, or is it more complex? Again, the authors found that they got the best performance by using a more complex approach, a Set2Set aggregator, which uses item-to-item attention within the set of final-layer atom representations to construct an aggregated grap-level embedding

The empirical tests within the paper highlight a few more interestingly relevant design choices that are less directly captured by the framework. The first is the fact that it's quite beneficial to explicitly include Hydrogen atoms as part of the graph, rather than just "attaching" them to their nearest-by atoms as a count that goes on that atom's feature vector. The second is that it's valuable to start out your edge features with a continuous representation of the spatial distance between atoms, along with an embedding of the bond type. 

This is particularly worth considering because getting spatial distance data for a molecule requires solving the free-energy problem to determine its spatial conformation, a costly process. We might ideally prefer a network that can work on bond information alone. The authors do find a non-spatial-information network that can perform reasonably well - reaching full accuracy on 5 of 13 targets, compared to 11 with spatial information. However, the difference is notable, which, at least from my perspective, begs the question of whether it'd ever be possible to learn representations that can match the performance of spatially-informed ones without explicitly providing that information.

This is an interesting - and refreshing - paper, in that, instead of trying to go all-in on a particular theoretical point, the authors instead run a battery of empirical investigations, all centered around the question of how to explain what happens to make transfer learning work. The experiments don't all line up to support a single point, but they do illustrate different interesting facets of the transfer process. 

- An initial experiment tries to understand how much of the performance of fine-tuned models can be explained by (higher-level, and thus larger-scale) features, and how much is driven by lower level (and thus smaller-scale) image statistics.  To start with, the authors compare the transfer performance from ImageNet onto three different datasets - clip art, sketches, and real images. As expected, transfer performance is highest with real datasets, which are the most similar to training domain. However, there still *is* positive transfer in terms of final performance across all domains, as well as benefit in optimization speed.
- To try to further tease out the difference between the transfer benefits of high and low-level features, the authors run an experiment where blocks of pixels are shuffled around within the image on downstream tasks .  The larger the size of the blocks being shuffled, the more that large-scale features of the image are preserved. As predicted, accuracy drops dramatically when pixel block size is small, for both randomly initialized and pretrained models. In addition, the relative value added by pretraining drops, for all datasets except quickdraw (the dataset of sketches). This suggests that in most datasets, the value brought by fine-tuning was mostly concentrated in large-scale features. One interesting tangent of this experiment was the examination of optimization speed (in the form of mean training accuracy over initial epochs). Even at block sizes too small for pretraining to offer a benefit to final accuracy, it did still contribute to faster training. (See transparent bars in right-hand plot below)
https://i.imgur.com/Y8sO1da.png

- On a somewhat different front, the authors look into how similar pretrained + finetuned models are to one another, compared to models trained on the same dataset from random initializations. First, they look at a measure of feature similarity, and find that the features learned by two pretrained networks are more similar to each other than a pretrained network is to a randomly initalized network, and also more than two randomly initialized networks are to one another. Randomly initialized networks are closest to one another in their final-layer features, but this is still a multiple of 4 or 5 less than the similarity between the pretrained networks
- Looking at things from the perspective of optimization, the paper measures how much performance drops when you linearly interpolate between different solutions found by both randomly initialized and pretrained networks. For randomly initialized networks, interpolation requires traversing a region where test accuracy drops to 0%. However, for pretrained networks, this isn't the case, with test accuracy staying high throughout. This suggests that pretraining gets networks into a basin of the loss landscape, and that future training stays within that basin. There were also some experiments on module criticality that I believe were in a similar vein to these, but which I didn't fully follow
- Finally, the paper looks at the relationship between accuracy on the original pretraining task and both accuracy and optimization speed on the downstream task. They find that higher original-task accuracy moves in the same direction as higher downstream-task accuracy, though this is less true when the downstream task is less related (as with quickdraw). Perhaps more interestingly, they find that the benefits of transfer to optimization speed happen and plateau quite early in training. Clip Art and Real transfer tasks are much more similar in the optimization speed benefits they get form ImageNet training, where on the accuracy front, the real did dramatically better.
https://i.imgur.com/jBCJcLc.png

While there's a lot to dig into in these results overall, the things I think are most interesting are the reinforcing of the idea that even very random and noisy pretraining can be beneficial to optimization speed (this seems reminiscent of another paper I read from this year's NeurIPS, examining why pretraining on random labels can help downstream training), and the observation that pretraining deposits weights in a low-loss bucket, from which they can learn more efficiently (though, perhaps, if the task is too divergent from the pretraining task, this difficulty in leaving the basin becomes a disadvantage). This feels consistent with some work in the Lottery Ticket Hypothesis, which has recently suggested that, after a short duration of training, you can rewind a network to a checkpoint saved after that duration, and be successfully able to train to low loss again.

This paper deals with the question what / how exactly CNNs learn, considering the fact that they usually have more trainable parameters than data points on which they are trained.

When the authors write "deep neural networks", they are talking about Inception V3, AlexNet and MLPs.

## Key contributions

* Deep neural networks easily fit random labels (achieving a training error of 0 and a test error which is just randomly guessing labels as expected). $\Rightarrow$Those architectures can simply brute-force memorize the training data.
* Deep neural networks fit random images (e.g. Gaussian noise) with 0 training error. The authors conclude that VC-dimension / Rademacher complexity, and uniform stability are bad explanations for generalization capabilities of neural networks
* The authors give a construction for a 2-layer network with $p = 2n+d$ parameters - where $n$ is the number of samples and $d$ is the dimension of each sample - which can easily fit any labeling. (Finite sample expressivity). See section 4.


## What I learned

* Any measure $m$ of the generalization capability of classifiers $H$ should take the percentage of corrupted labels ($p_c \in [0, 1]$, where $p_c =0$ is a perfect labeling and $p_c=1$ is totally random) into account: If $p_c = 1$, then $m()$ should be 0, too, as it is impossible to learn something meaningful with totally random labels.
* We seem to have built models which work well on image data in general, but not "natural" / meaningful images as we thought.


## Funny

> deep neural nets remain mysterious for many reasons

> Note that this is not exactly simple as the kernel matrix requires 30GB to store in memory. Nonetheless, this system can be solved in under 3 minutes in on a commodity workstation with 24 cores and 256 GB of RAM with a conventional LAPACK call.


## See also

* [Deep Nets Don't Learn Via Memorization](https://openreview.net/pdf?id=rJv6ZgHYg)

Bechtle et al. propose meta learning via learned loss ($ML^3$) and derive and empirically evaluate the framework on classification, regression, model-based and model-free reinforcement learning tasks. 

The problem is formalized as learning parameters $\Phi$ of a meta loss function $M_\phi$ that computes loss values $L_{learned} = M_{\Phi}(y, f_{\theta}(x))$. Following the outer-inner loop meta algorithm design the learned loss $L_{learned}$ is used to update the parameters of the learner in the inner loop via gradient descent:
$\theta_{new} = \theta - \alpha \nabla_{\theta}L_{learned} $. The key contribution of the paper is the way to construct a differentiable learning signal for the loss parameters $\Phi$.

The framework requires to specify a task loss $L_T$ during meta train time, which can be for example the mean squared error for regression tasks. After updating the model parameters to $\theta_{new}$ the task loss is used to measure how much learning progress has been made with loss parameters $\Phi$. The key insight is the decomposition via chain-rule of  $\nabla_{\Phi} L_T(y, f_{\theta_{new}})$:

$\nabla_{\Phi} L_T(y, f_{\theta_{new}}) = \nabla_f L_t \nabla_{\theta_{new}}f_{\theta_{new}} \nabla_{\Phi} \theta_{new} = \nabla_f L_t \nabla_{\theta_{new}}f_{\theta_{new}} [\theta - \alpha \nabla_{\theta} \mathbb{E}[M_{\Phi}(y, f_{\theta}(x))]]$.

This allows to update the loss parameters with gradient descent as: $\Phi_{new} = \Phi - \eta \nabla_{\Phi} L_T(y, f_{\theta_{new}})$. 

This update rules yield the following $ML^3$ algorithm for supervised learning tasks:

https://i.imgur.com/tSaTbg8.png

For reinforcement learning the task loss is the expected future reward of policies induced by the policy $\pi_{\theta}$, for model-based rl with respect to the approximate dynamics model and for the model free case a system independent surrogate: $L_T(\pi_{\theta_{new}}) = -\mathbb{E}_{\pi_{\theta_{new}}} \left[ R(\tau_{\theta_{new}})  \log \pi_{\theta_{new}}(\tau_{new})\right] $. 

The allows further to incorporate extra information via an additional loss term $L_{extra}$ and to consider the augmented task loss $\beta L_T + \gamma L_{extra} $, with weights $\beta, \gamma$ at train time. Possible extra loss terms are used to add physics priors, encouragement of exploratory behavior or to incorporate expert demonstrations. The experiments show that this, at test time unavailable information, is retained in the shape of the loss landscape. 

The paper is packed with insightful experiments and shows that the learned loss function:
- yields in regression and classification better accuracies at train and test tasks
- generalizes well and speeds up learning in model based rl tasks
- yields better generalization and faster learning in model free rl
- is agnostic across a bunch of evaluated architectures (2,3,4,5 layers)
- with incorporated extra knowledge yields better performance than without and is superior to alternative approaches like iLQR in a MountainCar task. 

The paper introduces a promising alternative, by learning the loss parameters, to MAML like approaches that learn the model parameters. It would be interesting to see if the learned loss function generalizes better than learned model parameters to a broader distribution of tasks like the meta-world tasks. 

This work attempts to use meta-learning to learn an update rule for a reinforcement learning agent. In this context, "learning an update rule" means learning the parameters of an LSTM module that takes in information about the agent's recent reward and current model and outputs two values - a scalar and a vector - that are used to update the agent's model. I'm not going to go too deep into meta-learning here, but, at a high level, meta learning methods optimize parameters governing an agent's learning, and, over the course of many training processes over many environments, optimize those parameters such that the reward over the full lifetime of training is higher. 

To be more concrete, the agent in a given environment learns two things: 

- A policy, that is, a distribution over predicted action given a state.
- A "prediction vector". This fits in the conceptual slot where most RL algorithms would learn some kind of value or Q function, to predict how much future reward can be expected from a given state. However, in this context, this vector is *very explicitly* not a value function, but is just a vector that the agent-model generates and updates. The notion here is that maybe our human-designed construction of a value function isn't actually the best quantity for an agent to be predicting, and, if we meta-learn, we might find something more optimal. I'm a little bit confused about the structure of this vector, but I think it's *intended* to be a categorical 1-of-m prediction

At each step, after acting in the environment, the agent passes to an LSTM: 

- The reward at the step
- A binary of whether the trajectory is done
- The discount factor
- The probability of the action that was taken from state t
- The prediction vector evaluated at state t
- The prediction vector evaluated at state t+1

Given that as input (and given access to its past history from earlier in the training process), the LSTM predicts two things: 

- A scalar, pi-hat
- A prediction vector, y-hat

These two quantities are used to update the existing policy and prediction model according to the rule below.

https://i.imgur.com/xx1W9SU.png

 Conceptually, the scalar governs whether to increase or decrease probability assigned to the taken action under the policy, and y-hat serves as a target for the prediction vector to be pulled towards.  An important thing to note about the LSTM structure is that none of the quantities it takes as input are dependent on the action or observation space of the environment, so, once it is learned it can (hopefully) generalize to new environments. 

Given this, the basic meta learning objective falls out fairly easily - optimize the parameters of the LSTM to maximize lifetime reward, taken in expectation over training runs.  However, things don't turn out to be quite that easy. The simplest version of this meta-learning objective is wildly unstable and difficult to optimize, and the authors had to add a number of training hacks in order to get something that would work. (It really is dramatic, by the way, how absolutely essential these are to training something that actually learns a prediction vector). These include: 

- A entropy bonus, pushing the meta learned parameters to learn policies and prediction vectors that have higher entropy (which is to say: are less deterministic)
- An L2 penalty on both pi-hat and y-hat
- A removal of the softmax that had originally been originally taken over the k-dimensional prediction vector categorical, and switching that target from a KL divergence to a straight mean squared error loss. As far as I can tell, this makes the prediction vector no longer actually a 1-of-k categorical, but instead just a continuous vector, with each value between 0 and 1, which makes it make more sense to think of k separate binaries? This I was definitely confused about in the paper overall

https://i.imgur.com/EL8R1yd.png

With the help of all of these regularizers, the authors were able to get something that trained, and that appeared to be able to perform comparably to or better than A2C - the human-designed baseline - across the simple grid-worlds it was being trained in. However, the two most interesting aspects of the evaluation were: 

1. The authors showed that, given the values of the prediction vector, you could predict the true value of a state quite well, suggesting that the vector captured most of the information about what states were high value. However, beyond that, they found that the meta-learned vector was able to be used to predict the value calculated with discount rates different that than one used in the meta-learned training, which the hand-engineered alternative, TD-lambda, wasn't able to do (it could only well-predict values at the same discount rate used to calculate it). This suggests that the network really is learning some more robust notion of value that isn't tied to a specific discount rate. 

2. They also found that they were able to deploy the LSTM update rule learned on grid worlds to Atari games, and have it perform reasonably well - beating A2C in a few cases, though certainly not all. This is fairly impressive, since it's an example of a rule learned on a different, much simpler set of environments generalizing to more complex ones, and suggests that there's something intrinsic to Reinforcement Learning that it's capturing