TLDR; The authors propose Progressive Neural Networks (ProgNN), a new way to do transfer learning without forgetting prior knowledge (as is done in finetuning). ProgNNs train a neural neural on task 1, freeze the parameters, and then train a new network on task 2 while introducing lateral connections and adapter functions from network 1 to network 2. This process can be repeated with further columns (networks). The authors evaluate ProgNNs on 3 RL tasks and find that they outperform finetuning-based approaches.

#### Key Points

- Finetuning is a destructive process that forgets previous knowledge. We don't want that.
- Layer h_k in network 3 gets additional lateral connections from layers h_(k-1) in network 2 and network 1. Parameters of those connections are learned, but network 2 and network 1 are frozen during training of network 3.
- Downside: # of Parameters grows quadratically with the number of tasks. Paper discussed some approaches to address the problem, but not sure how well these work in practice.
- Metric: AUC (Average score per episode during training) as opposed to final score. Transfer score = Relative performance compared with single net baseline. 
- Authors use Average Perturbation Sensitivity (APS) and Average Fisher Sensitivity (AFS) to analyze which features/layers from previous networks are actually used in the newly trained network.
- Experiment 1: Variations of Pong game. Baseline that finetunes only final layer fails to learn. ProgNN beats other  baselines and APS shows re-use of knowledge.
- Experiment 2: Different Atari games. ProgNets result in positive Transfer 8/12 times, negative transfer 2/12 times. Negative transfer may be a result of optimization problems. Finetuning final layers fails again. ProgNN beats other approaches.
- Experiment 3: Labyrinth, 3D Maze. Pretty much same result as other experiments.


#### Notes

- It seems like the assumption is that layer k always wants to transfer knowledge from layer (k-1). But why is that true? Network are trained on different tasks, so the layer representations, or even numbers of layers, may be completely different. And Once you introduce lateral connections from all layers to all other layers the approach no longer scales.
- Old tasks cannot learn from new tasks. Unlike humans.
- Gating or residuals for lateral connection could make sense to allow to network to "easily" re-use previously learned knowledge.
- Why use AUC metric? I also would've liked to see the final score. Maybe there's a good reason for this, but the paper doesn't explain.
- Scary that finetuning the final layer only fails in most experiments. That's a very commonly used approach in non-RL domains.
- Someone should try this on non-RL tasks.
- What happens to training time and optimization difficult as you add more columns? Seems prohibitively expensive.

This paper describes using Relation Networks (RN) for reasoning about relations between objects/entities.
RN is a plug-and-play module and although expects object representations as input,
the semantics of what an object is need not be specified, so object representations
can be convolutional layer feature vectors or entity embeddings from text, or something else.
And the feedforward network is free to discover relations between objects (as opposed to being
hand-assigned specific relations).

- At its core, RN has two parts:
	- a feedforward network `g` that operates on pairs of object representations,
	for all possible pairs, all pairwise computations pooled via element-wise addition
	- a feedforward network `f` that operates on pooled features for downstream
	task, everything being trained end-to-end

- When dealing with pixels (as in CLEVR experiment), individual object representations are
spatially distinct convolutional layer features (196 512-d object representations for VGG conv5 say).
The other experiment on CLEVR uses explicit factored object state representations with 3D coordinates,
shape, material, color, size.

- For bAbI, object representations are LSTM encodings of supporting sentences.

- For VQA tasks, `g` conditions its processing on question encoding as well, as relations
that are relevant for figuring out the answer would be question-dependent.


## Strengths

- Very simple idea, clearly explained, performs well. Somewhat shocked that it
hasn't been tried before.

## Weaknesses / Notes

Fairly simple idea — let a feedforward network
operate on all pairs of object representations and figure out relations
necessary for downstream task with end-to-end training. And it is fairly general in its design,
relations aren't hand-designed and neither are object representations — for
RGB images, these are spatially distinct convolutional layer features, for text,
these are LSTM encodings of supporting facts, and so on. This module can be dropped
in and combined with more sophisticated networks to improve performance at VQA.

RNs also offer an alternative design choice to prior works on CLEVR, that have
this explicit notion of programs or modules with specialized roles (that need to be pre-defined),
as opposed to letting these relations emerge, reducing dependency on hand-designing
modules and adding in inductive biases from an architectural point-of-view for
the network to reason about relations (earlier end-to-end VQA models didn't have
the capacity to figure out relations).

This paper is about a recommendation system approach using collaborative filtering (CF) on implicit feedback datasets.

The core of it is the minimization problem

$$\min_{x_*, y_*} \sum_{u,i} c_{ui} (p_{ui} - x_u^T y_i)^2 + \underbrace{\lambda \left ( \sum_u || x_u ||^2 + \sum_i || y_i ||^2\right )}_{\text{Regularization}}$$

with

* $\lambda \in [0, \infty[$ is a hyper parameter which defines how strong the model is regularized
* $u$ denoting a user, $u_*$ are all user factors $x_u$ combined
* $i$ denoting an item, $y_*$ are all item factors $y_i$ combined
* $x_u \in \mathbb{R}^n$ is the latent user factor (embedding); $n$ is another hyper parameter. $n=50$ seems to be a reasonable choice.
* $y_i \in \mathbb{R}^n$ is the latent item factor (embedding)
* $r_{ui}$ defines the "intensity"; higher values mean user $u$ interacted more with item $i$
* $p_{ui} = \begin{cases}1 & \text{if } r_{ui} >0\\0 &\text{otherwise}\end{cases}$
* $c_{ui} := 1 + \alpha r_{ui}$ where $\alpha \in [0, \infty[$ is a hyper parameter; $\alpha =40$ seems to be reasonable

In contrast, the standard matrix factoriation optimization function looks like this ([example](https://www.cs.cmu.edu/~mgormley/courses/10601-s17/slides/lecture25-mf.pdf)):

$$\min_{x_*, y_*} \sum_{(u, i, r_{ui}) \in \mathcal{R}} {(r_{ui} - x_u^T y_i)}^2  + \underbrace{\lambda \left ( \sum_u || x_u ||^2 + \sum_i || y_i ||^2\right )}_{\text{Regularization}}$$

where

* $\mathcal{R}$ is the set of all ratings $(u, i, r_{ui})$ - user $u$ has rated item $i$ with value $r_{ui} \in \mathbb{R}$

They use alternating least squares (ALS) to train this model.

The prediction then is the dot product between the user factor and all item factors ([source](https://github.com/benfred/implicit/blob/master/implicit/recommender_base.pyx#L157-L176))

We want to find two matrices $W$ and $H$ such that $V = WH$. Often a goal is to determine underlying patterns in the relationships between the concepts represented by each row and column. $W$ is some $m$ by $n$ matrix and we want the inner dimension of the factorization to be $r$. So 

$$\underbrace{V}_{m \times n} = \underbrace{W}_{m \times r} \underbrace{H}_{r \times n}$$

Let's consider an example matrix where of three customers (as rows) are associated with three movies (the columns) by a rating value.

$$
V = \left[\begin{array}{c c c}
5 & 4 & 1  \\\\
4 & 5 & 1 \\\\
2 & 1 & 5
\end{array}\right]
$$


We can decompose this into two matrices with $r = 1$. First lets do this without any non-negative constraint using an SVD reshaping matrices based on removing eigenvalues:


$$
W = \left[\begin{array}{c c c}
-0.656 \\\
 -0.652 \\\
 -0.379
\end{array}\right],
H = \left[\begin{array}{c c c}
-6.48 & -6.26 & -3.20\\\\
\end{array}\right]
$$

We can also decompose this into two matrices with $r = 1$ subject to the constraint that $w_{ij} \ge 0$ and  $h_{ij} \ge 0$. (Note: this is only possible when $v_{ij} \ge 0$):

$$
W = \left[\begin{array}{c c c}
0.388 \\\\
0.386 \\\\
0.224
\end{array}\right],
H = \left[\begin{array}{c c c}
11.22 & 10.57 & 5.41  \\\\
\end{array}\right]
$$

Both of these $r=1$ factorizations reconstruct matrix $V$ with the same error. 

$$
V \approx WH = \left[\begin{array}{c c c}
4.36 & 4.11 & 2.10 \\\
4.33 & 4.08 & 2.09 \\\
2.52 & 2.37 & 1.21 \\\
\end{array}\right]
$$


If they both yield the same reconstruction error then why is a non-negativity constraint useful? We can see above that it is easy to observe patterns in both factorizations such as similar customers and similar movies. `TODO: motivate why NMF is better`



#### Paper Contribution 

This paper discusses two approaches for iteratively creating a non-negative $W$ and $H$ based on random initial matrices. The paper discusses a multiplicative update rule where the elements of $W$ and $H$ are iteratively transformed by scaling each value such that error is not increased. 

The multiplicative approach is discussed in contrast to an additive gradient decent based approach where small corrections are iteratively applied. The multiplicative approach can be reduced to this by setting the learning rate ($\eta$) to a ratio that represents the magnitude of the element in $H$ to the scaling factor of $W$ on $H$.



### Still a draft

This is follow-up work to the ResNets paper. It studies the propagation formulations behind the connections of deep residual networks and performs ablation experiments. A residual block can be represented with the equations $y_l = h(x_l) + F(x_l, W_l); x_{l+1} = f(y_l)$. $x_l$ is the input to the l-th unit and $x_{l+1}$ is the output of the l-th unit. In the original ResNets paper, $h(x_l) = x_l$, $f$ is ReLu, and F consists of 2-3 convolutional layers (bottleneck architecture) with BN and ReLU in between. In this paper, they propose a residual block with both $h(x)$ and $f(x)$ as identity mappings, which trains faster and performs better than their earlier baseline. Main contributions:

- Identity skip connections work much better than other multiplicative interactions that they experiment with:
    - Scaling $(h(x) = \lambda x)$: Gradients can explode or vanish depending on whether modulating scalar \lambda > 1 or < 1.
    - Gating ($1-g(x)$ for skip connection and $g(x)$ for function F):
    For gradients to propagate freely, $g(x)$ should approach 1, but
    F gets suppressed, hence suboptimal. This is similar to highway
    networks. $g(x)$ is a 1x1 convolutional layer.
    - Gating (shortcut-only): Setting high biases pushes initial $g(x)$
    towards identity mapping, and test error is much closer to baseline.
    - 1x1 convolutional shortcut: These work well for shallower networks
    (~34 layers), but training error becomes high for deeper networks,
    probably because they impede gradient propagation.

- Experiments on activations.
    - BN after addition messes up information flow, and performs considerably
    worse.
    - ReLU before addition forces the signal to be non-negative, so the signal is monotonically increasing, while ideally a residual function should be free to take values in (-inf, inf).
    - BN + ReLU pre-activation works best. This also prevents overfitting, due
    to BN's regularizing effect. Input signals to all weight layers are normalized.

## Strengths

- Thorough set of experiments to show that identity shortcut connections
are easiest for the network to learn. Activation of any deeper unit can
be written as the sum of the activation of a shallower unit and a residual
function. This also implies that gradients can be directly propagated to
shallower units. This is in contrast to usual feedforward networks, where
gradients are essentially a series of matrix-vector products, that may vanish, as networks grow deeper.

- Improved accuracies than their previous ResNets paper.

## Weaknesses / Notes

- Residual units are useful and share the same core idea that worked in
LSTM units. Even though stacked non-linear layers are capable of asymptotically
approximating any arbitrary function, it is clear from recent work that
residual functions are much easier to approximate than the complete function.
The [latest Inception paper](http://arxiv.org/abs/1602.07261) also reports
that training is accelerated and performance is improved by using identity
skip connections across Inception modules.

- It seems like the degradation problem, which serves as motivation for
residual units, exists in the first place for non-idempotent activation
functions such as sigmoid, hyperbolic tan. This merits further
investigation, especially with recent work on function-preserving transformations such as [Network Morphism](http://arxiv.org/abs/1603.01670), which expands the Net2Net idea to sigmoid, tanh, by using parameterized activations, initialized to identity mappings.